The Curious Case of State Economies in India

Data v OpinionThe political economy debate in India currently has all kinds of battle lines drawn. Market versus Social, Social versus Mixed, Central versus Federal – you name it and the combination of warring parties exist. Political leaders cutting across state and party lines never fail to exploit an opportunity to show how wonderful their corner is – often taking recourse to data (which is a good thing, as the opening quote for this piece clearly demonstrates). The trend for debates – and hopefully journalism too at some time – getting removed from passion and becoming increasingly data centric bodes very well for a vast and diverse country like India. However, for any conclusion drawn from data analysis to be worthwhile a sine-qua-non criteria is definitely acceptable data quality.  Toying around with public data released by various government and quasi government agencies in India, it is alarming to discover that several key indicators, indicators that routinely weave their way in our debates and more importantly – for shaping policy decisions, are rather suspect in data quality

Take a very basic indicator like GDP for instance. This is one bellwether that investors, policy makers, political guardians – everyone – loves to brandish (when the going is good) as a measure of achievement. It is not unusual for representatives from Indian states espousing how theirs grew at a faster clip than another – or for that matter grew faster than the country as a whole. How solid are these claims? Taking data from the Central Statistical Organization, a very rudimentary analysis shows what I’d like to call the curious case of Indian State GDP

Sum of state GDP value NEVER totals the India number. India level numbers are always higher. State GDP total varianceThere is some benevolent God amongst the 300 million worshipped in the country that makes good whatever is the shortfall is to our collective aspirations. The difference was showing a declining trend until the last year came about and created a record high!

And then there is the case with GDP growth rates too. Between 2004-05 to 2012-13, growth rates calculated by summing state GDP has been higher than the India growth rate on three occasions. However, looking at just the top ten states (by GDP size), in the same period the growth rate of at least five of the ten exceeded the overall growth rate of the country. In 2012-13, nine of the top ten reported higher growth rates that what India reported! As you can see from the graph below, other than in 2007 and 2008 when our statisticians got very efficient (or extraordinarily lucky), the difference between calculated GDP growth rate and that actually announced has been quite substantialState GDP growth variance

Other than structural inconsistency, there are wild swing in individual numbers as well. Consider Karnataka, my home state. After clocking a 7.11% growth in 2008-09 (India 6.72%) it plummeted to a mere 1.30% in 2009-10 (India 8.59%) only to rise like a phoenix and deliver a jaw dropping 10.15% in 2010-11 (India 8.91%). Amongst the top ten states Uttar Pradesh and West Bengal show the lowest standard deviation of the GDP growth numbers between 2004-05 to 2012-13

In conclusion we go back to the point we started with. A political or economic discourse, leave alone policy making, cannot be based on data that is this fragile. And this is just one set of cross verifiable data that I have looked at – heaven knows what lurks behind other datasets. As Indian polity becomes modern and policy making gets data-centric, it is imperative that the government increases the volume and velocity of open data and provides (possibly in partnership with the private sector) solid data toolsets for everyone to access, qualify and use the outcome to drive debates. Only through a process of many eyes looking at data, critiques on veracity and wider usage that the quality of information shall improve. Else, extending Jim Barksdale’s suggestion, we might as well go with the most handsome guy’s opinion