Home / Opinion / Columns /  The accidental statistician who taught world to estimate farm output

For most of the British Raj, India’s crop estimates were based on data from local revenue officials. The Bengal famine of 1943 led to severe criticism of such estimates. The British administration turned to the statistical adviser at the Imperial (later Indian) Council of Agricultural Research (ICAR), Pandurang Vasudeo Sukhatme for help.

Sukhatme proposed supplementing the administrative structure for collecting data with modern sampling methods. With some training in statistical methods and supervision by trained staff, the local land revenue official could collect reliable data on yields, he argued. Thus was the idea of sampling random plots to quickly estimate yields implemented for the first time in India. Along with his trusted colleague, Vinayak Govind Panse, Sukhatme began training agricultural staff in survey methods. The duo also experimented with different plot sizes to arrive at the optimal size for estimates. Sukhatme and Panse were among the first to write books on sampling methods based on their field experiments, and these came to be used by statisticians globally. As the head of the statistics division of the Food and Agriculture Organization (FAO), Sukhatme helped modernize farm data collection across the world.

If P.C. Mahalanobis taught the world how to study household behaviour through random sampling, Sukhatme taught it how to estimate agricultural output through a mix of survey and census methods. Both were statistical pioneers, but had their share of differences. Mahalanobis had made his name in estimating crop yields in Bengal, and found the younger statistician’s recommended plot sizes too large and his reliance on local administrative sources problematic; he proposed that the National Sample Survey (NSS) team he was setting up at the Indian Statistical Institute (ISI) be entrusted with the task of estimating crop yields.

Sukhatme held his ground, arguing that the ‘ad hoc’ NSS enumerators employed by Mahalanobis were likely to be unfamiliar with local conditions. Local revenue officials were also well placed to observe crops over their entire life-cycle. NSS enumerators would inevitably pick the plots with ripe crops at the time of their visit, and this would lead to biased (non-random) sampling, he argued. Given that local revenue officials were already familiar with the concept of large rectangular plots, Sukhatme found these ideal in collecting data on yields.

Studies conducted by a joint ICAR-ISI team and later by the Planning Commission suggested that both methods gave broadly similar results. Yet, the controversy persisted, with Mahalanobis’ successors (at ISI) and Sukhatme’s followers (at ICAR) insisting on following their ‘own’ method.

In an obituary of Sukhatme published by the Economic and Political Weekly in February 1997, the Pune-based statistician Anil Gore wrote that Sukhatme’s differences with Mahalanobis may have led to his exit from the ICAR. When Sukhatme’s plans to develop a separate institute for agricultural statistics appeared to be in jeopardy because of opposition from Mahalanobis, the then agriculture minister K M Munshi advised him to take up the FAO assignment, Gore wrote.

In a 1999 research paper on the evolution of India’s statistical system, the ISI statisticians J.K. Ghosh, P. Maiti, T.J. Rao and B.K. Sinha argued that the differing origins of the two statisticians could have played a role in Independent India’s first major statistical controversy. Sukhatme hailed from a region where the patwari played a prominent role. In Bengal, this post had been abolished by the British. “What appeared to be a scientific controversy was rooted in the social background of the scientists," Ghosh and his co-authors wrote.

While their social backgrounds differed, Sukhatme and Mahalanobis were both accidental statisticians. In the former’s case, it was literally an accident that made him pursue statistics. After graduating from Fergusson College in Pune, Sukhatme wanted to appear for the Indian Civil Service examination. On the way to the exam in London, he had an accident. Rather than wait for the next attempt, he enrolled in a course in statistics in the University College London. Sukhatme would go on to study under the world’s leading statisticians, such as J. Neyman and R.A. Fisher, before returning to India.

He briefly worked as an epidemiologist in Kolkata before going on to head the statistics wing at the ICAR, and later the FAO. While at the FAO, Sukhatme’s work on the measurement of under-nutrition won him the prestigious Guy Medal of the Royal Statistical Society. His research suggested that calorie requirements varied sharply across individuals with similar lifestyles, and it varied from day to day even for the same individual. The Sukhatme-Margen hypothesis—which suggests that at low calorie intake levels, stored energy in the body is used with greater metabolic efficiency—is named after him and his co-author, nutritionist Sheldon Margen.

His views on nutrition led him to criticize the calorie-based poverty norms proposed by another Punekar, V.M. Dandekar. Dandekar’s proposal won official acceptance but Sukhatme’s critique of calorie-based poverty norms continues to resonate. His ideas on blending survey and administrative data sources find more takers today. Like Mahalanobis, Sukhatme believed statisticians should be familiar with the subject where they apply their tools. This idea too retains its relevance.

This is the second of a five-part series on the founders of India’s once-renowned statistical system.

Pramit Bhattacharya is a Chennai-based journalist. His Twitter handle is pramit_b

Catch all the Business News, Market News, Breaking News Events and Latest News Updates on Live Mint. Download The Mint News App to get Daily Market Updates.
More Less
Subscribe to Mint Newsletters
* Enter a valid email
* Thank you for subscribing to our newsletter.

Recommended For You

Get alerts on WhatsApp
Set Preferences My ReadsWatchlistFeedbackRedeem a Gift CardLogout