India's Central Statistical Office (CSO) is working hard to modernise India's data collection methods and also the way it calculates various economic indices. A couple of years ago, it had changed the way GDP was calculated.
Recently, it has changed both the base year and the composition of both the Index of Industrial Production (IIP) as well as the wholesale price index (WPI). The base year for both indices were changed to 2011-12, against the 2004-05 earlier. Both were sorely needed because other indices (including the GDP and consumer price index (CPI) had already started using 2011-12 as the base year). Also, a number of goods that had no relevance were changed in the IIP, while adding new goods which make more sense. Ditto for the WPI, which has seen a change in its basket of goods.
The changes to IIP and WPI could further affect the GDP calculations, though the immediate impact is hard to gauge. What we do know is that once the Goods & Services Tax (GST) starts rolling, there will be even more data available, and that might further lead to changes in the way GDP is calculated. In fact, it could completely change the way GDP is measured in the country, as there will be a seamless, near real-time data available on goods and services supplied and sold. There is also the plan to replace the IIP eventually with a new index that measures services as well, and not merely manufacturing , mining and electricity production.
The CSO's efforts are to make most indices more in line with what developed nations around the world measure. These were much needed.
The CSO revisions have raised some controversy - especially the GDP calculations because the back series is not being published and because it bumped up growth in the immediate past. But what is important is that the new method tries to capture more data and more relevant data.
However, there are other areas where our statistics - both the collection methods as well as the calculation methodology - needs drastic overhaul. These are crucial areas such as jobs created, number of unemployed and under employed, farmer incomes etc.
Take job creation data for example. What we currently get is a patchwork of data, which does not in any way reflect the real picture. For example, the ministry of labour comes out with a quarterly survey of jobs created/jobs lost, but that covers all of 8 sectors, and is fairly limited in its scope. It doesn't capture the other, fast growing sectors or even fully represent the services sector, which constitutes the biggest chunk of the economy. It also does not capture the labour movements in the informal sector, which is estimated to account for over 70 per cent of the total employed. National Sample Survey Organisation (NSSO) data on employment is more indepth and presents a far more comprehensive picture - but it is five years out of date in most cases. Between the survey and the time the data is turned into a report, multiple things change and therefore the NSSO data is not great for policy making when jobs fluctuate not just year to year, but also season to season.
Ditto of other economic and socio economic data. The prime minister talked of doubling farmer's incomes within five years recently. The problem is that no one is quite clear what should be taken as the current income levels for the average farmer given that the latest data is of 2013-14.The NSSO also surveys a number of other important things, including spending patterns and consumption habits of the poor etc which could be of crucial help for policy making only if the data did not come out so late.
What is needed is for a total overhaul of India's data collection and analysis methods. Luckily, the rapid advances of technology could make both surveying (using mobile applications etc for capturing data etc) and analysing data quicker much more feasible. What we need is robust data on all sorts of subjects from a reliable source - from jobs being created, to taxes being paid, houses being built, production of different industries etc. And for that, the way we have traditionally collected and collated data needs total rethink.