The solution isn’t to have lots of small surveys done on different parts of India and assume that data from each is representative of all of India.
The Reserve Bank of India governor has expressed the urgent need for better official data on which to base policy, and pointed out the dangers of basing policy on incorrect or unstable data, but the ministry of statistics and programme implementation has been given to someone who reportedly doesn’t want this unimportant job. Yet this function does need far more energy and participation from a wider set of players, particularly when the country is going through rapid change that is so unstructured. As the heterogeneity of the new India grows, there is a greater need for more mega sample size “ground-level” surveys of households and individuals; and more micro (one- two- or three-person) informal enterprises to understand what is going on. The fundamental assumptions based on which the traditional, official and national surveys have been designed need to be re-visited. How India works and earns has gone through a great deal of change. There are many more Indias now than before, and their patterns of living, earning and spending are diverging. Several new kinds of multi-occupation individuals and households, and multi-earner households have emerged. Expenditure heads have undergone big changes too. We probably don’t know enough to even know what the changes are, and hence what needs to be changed in the design and questions and the measures in the surveys that are traditionally done. State-level disparities are far wider than ever before, and the state is becoming the unit of both policy action and business planning. Anecdotally, much of this is known. But analytical rigour to these anecdotes is still to come. Besides the anecdotes, all true, often contradict each other depending on where they are drawn from. We need to go through a discovery phase from scratch. And discovery requires many more discoverers, each looking at the same population with different lenses. We have a hang up about “official” and “not official” data bases, rather than thinking in terms of “good” and “bad” data bases. We have a horror of different data bases not “matching.” But its time to shed that horror — more people trying to solve a problem using different tools and methodologies, and even more people working on explanations for the divergences will eventually cause greater clarity in the form of newer and more robust mental models of current trends. Earlier, expenditure was a good surrogate of income. Now it isn’t for the middle and upper income groups. Earlier rural households were easily classifiable into where their main source of income came from — agriculture or not; now it is more complex than that. Earlier the categories of expenditure for the rich and the poor were pretty much the same. Now they are not. We now need to measure indebtedness of the middle and upper class households a lot more carefully, to understand more than what the aggregate credit data of banks will tell us. Most of India does not have regular jobs. Yet everyone, for the most part, earns a livelihood doing something or the other. The services economy also includes all the micro providers of services of some kind or the other. We still don’t exactly know how India earns, leave alone spends based on the categorisations people use to think about their expenditure. If household expenditure is supposed to be a pillar of India’s economic growth, then the more we know about this the better; and one National Sample Survey (NSS) alone isn’t enough to capture such a hydra-headed monster that is at least three countries in one. The solution isn’t to have lots of small surveys done on different parts of India and assume that data from each is representative of all of India; or to cobble them all together and say we have the whole jigsaw pieced together. The trouble is that we don’t even know what the whole jigsaw is supposed to be. How can it be? The need of the hour is to have large, descriptive surveys that describe thoroughly what is going on in India, grossed up from the household or the individual. Properly done, this requires humongous sample sizes, even with the smartest design. Who pays for this? The good news is that both corporate India and the policy-making India actually require pretty much the same data platforms, and for once, are willing to come together and volunteer time, advice and money for a good cause. There are no divergent data interests here. What diverges is how they use it. What analytics they bring to it, what conclusions they draw from it are to each, his own. Policymakers look at using data to influence outcomes, corporate India is agnostic to outcomes but looks at exploiting emergent patterns. To seriously make financial inclusion work by building sensible and safe business models, there needs to be a descriptive data base that is consumer-centred for everyone in the game. Build the data base, and the analytics will automatically follow, the same way apps are written prolifically to ride on tech platforms. Economists can look at the data using their tools and market strategists using theirs. Policy research organisations and think tanks must see development of such fundamental data bases as a core activity and shed the notion that this is not “pure play” policy, just low-end survey work. In order to walk down the policy development road, the road must be built first. And we don’t have such a road. And borrowing international comparisons as a substitute is unlikely to work either, because nowhere else in the world does such a confounding hybrid and “off pattern” evolution path exist.
The author is an independent market strategy consultant BS
