The power of narratives

In India, we all have heard the story behind Panchtantra. A wise king, despairing about the illiteracy of his heirs, commissions Vishnu Sharma to educate them. Sharma uses simple tales, each ending with a moral, and educates the princes in the art of governance. The advantage of a good story is that it holds the listeners’ attention and captures their minds and imagination. It is no surprise that narratives have such a powerful hold in modern day social science inquiry. 

In contrast to narratives, facts and fact-based presentations are messy, often delayed, and convey many conflicting messages. While a story can be told at any time, a fact-based analysis must necessarily await the compilation and analysis of data. In the case of the narrative, all that matters are the name and eloquence of the storyteller. Factual analysis, in contrast, must contend with all the limitations of data. The problem is that when a detailed analysis comes around to contest a narrative, the natural response is not to question the narrative but instead question the facts. There is a well-known example of this in the History of Science — the fabled existence of Luminiferous Ether. A series of careful experiments had challenged this narrative, but only when Albert Einstein formulated his Theory of Relativity years later, was the idea finally discarded. 

A more proximate example of this phenomenon is the recent controversy on the Gross Domestic Product (GDP) calculations from 2004 to 2011 period done by the Central Statistics Office (CSO). Let us examine some of the facts that underline this debate. It is well known that data pertaining to the informal or household sector in India is limited and often available with long time lags. The principal sources of data for the household sector in India are National Sample Survey Office's (NSSO) sample surveys of household and establishments, usually done at an interval of five years. In this case, NSSO had done household surveys of employment in 2004-05 and 2011-12, and surveys of establishments had been done in nearby periods. What emerged from these were some very striking facts. One, which was was widely noted by experts and commentators, was that between 2004-05 and 2011-12, we had one of the slowest growth rates in labour force. Its implications, however, for India’s GDP computation were not equally appreciated by most. Since the GDP base was revised to 1999-2000, the CSO would compute the benchmark estimate for value added in the household or informal sector by combining labour force estimates from household surveys with labour productivity estimates from establishment surveys. The benchmark estimate would then be projected in non-survey years by indicators that were more regularly available, or by simply extending the growth profile from earlier benchmark calculations. Slow growth in labour force means that under this method, the value added in 2011-12 had been over-estimated. The extent of over-statement would have been almost 100 per cent. In other words, the actual value added in the informal sector was almost half of what had originally been computed. A fact that most critics on the new GDP series conveniently ignore is that, if the methodology and other data sources had not been revised, we would have needed to revise our GDP estimate for 2011-12 downwards by almost 25 per cent. Its consequential impact on the growth between 2004-05 and 2011-12 can be similarly evaluated. 

The magnitude of this effect was masked by all the other changes in data and methodology done during the 2011-12 revision that capped the overall restatement to only about 3 per cent. However, behind this modest reduction in aggregate Gross Value Added (GVA) were sharp compositional changes in value added with significant reductions in the contribution of the household sector (and within it for retail trade), and significant increases in the share of the corporate sector (and within it for manufacturing). These compositional changes underlie the changes in growth pattern in current and constant prices, as well as the underlying structure of price deflators. These changes paraphrase Sherlock Holmes’ famous dictum “when you have eliminated the impossible, whatever remains, however improbable, must be the truth”. Unfortunately, the factual story obscures the narrative of “India Shining” and thus must be denied.

The problem is that factual analysis comes too late, and the requirements of political discourse in a vibrant democracy requires immediate narratives. 

A similar exercise is therefore underway in the construction of narratives on the impact of demonetisation and goods and services tax implementation. The factual data for the period is for the most part still coming in and unfolding, but the contesting and contrasting narratives have already gained their adherents. The chances are that when a fact-based analysis becomes available, it will please none of these disputants, and they will likely challenge the facts. At this stage, one can only appeal to the social scientists to remember and adhere to what is perhaps Holmes’ most significant advice to us “it is a capital mistake to theorise before you have all the evidence. It biases the judgment.”

The writer is former Chief Statistician of India