The Reserve Bank’s celebrated the ‘10th Annual Statistics Day Conference’ on July 26, 2016. The theme of the Conference was Risk Management and Big Data. Statistics Day is observed in India to commemorate the notable contributions made by Late Prof. Prasanta Chandra Mahalanobis in the fields of statistics, economic planning and development of the Indian statistical system. The Reserve Bank invites distinguished academicians and practitioners in the annual Statistics Day Conference to share their work and thoughts on a topical issue in the field of statistics, which is of special interest for the financial sector. The deliberations in the conference provide guidance for statistical work in the Bank.

Prof. Lars Peter Hansen, David Rockefeller Distinguished Service Professor of Economics at the University of Chicago and winner of the Nobel Memorial Prize in Economic Sciences in 2013 delivered the keynote address on Ambiguity Aversion and Model Misspecification. He said that statistics has dual role in constructing a dynamic economic model, where researchers depict economic actors coping with uncertainty and deduce the consequences inside a model, whereas unknown parameter estimation and model implications are assessed outside the model. There are clear statistical challenges in modelling uncertain investors, rational expectations and risk as there are limits to learning and drawing inferences, scope for behavioural distortions and statistical-uncertainty-induced fluctuations in prices of uncertainty. Modelling the broader perspective on uncertainty would involve modelling risk, ambiguity and changes over time. He said that misspecification is a challenge in statistical models used in practice and there are advances in decision theory that give statisticians structured ways to frame evidence to provide support to policy makers. Concerns about model misspecification in dynamic settings often feature long-term uncertainty and ambiguity aversion targets the forms of misspecification with the most adverse consequences. Prof. Hansen emphasised that one needs to acknowledge limits to human understanding of the linkages between financial markets and the macro-economy, where simple robust policies avoid adding uncertainty to the economic environment.

Dr. Alberto Cavallo, Associate Professor of Applied Economics at the Massachusetts Institute of Technology (MIT) Sloan School of Management and a Faculty Research Fellow at the NBER delivered a Special Lecture on The Billion Prices Project: Using Online Prices for Measurement and Research. He described the idea behind the Billion Prices Project at MIT, which is an academic initiative that pioneered the use of online data to conduct research on high-frequency price dynamics and inflation measurement. Under the Billion Prices Project, daily data are being collected since 2008 from large multi-channel retailers and the coverage has increased to over 60 countries. He dwelt upon the initial pilot project in Argentina as a validation exercise and its quick spread to cover other countries. He detailed the information collection mechanism and comparison with other micro-price data sources (viz., CPI and Scanner data). He also reflected upon the issues relating to quality changes, inadequate coverage of services, low representativeness, better timeliness and faster adjustment of relative prices in on-line data where product entry/exit are captured more rapidly than traditional price indices. He also shared the results of empirical investigations on divergence between on-line and off-line prices of same products where price-levels were found identical for 70 per cent of the time but heterogeneity was witnessed at country-level and retailer-level. He said that this is one example where Big Data has provided measurement opportunity and better availability for macroeconomic analysis and research.

Dr. David Bholat, an Analyst in the Statistics and Regulatory Data Division within the Bank of England and a visiting fellow at the Newcastle University delivered a special lecture on Big Data in a Small World. He described the use of Big Data analytics in specific areas by the Bank of England for enabling efficient policy decisions. Big Data could be used to analyse emerging macroeconomic issues ranging from asset quality of banks to the threat of deglobalisation. He described the suitability of text data, social media feeds and granular-level data as new-age enabler for advanced analytics and data visualisation.

The Panel Discussion on Application of Big Data and its prospects going forward was chaired by Dr. R.B. Barman, Chairman, National Statistical Commission, Government of India. The other panellists included Prof. D Ramakrishnan of Indian Institute of Technology, Mumbai, who is also the National Coordinator of the Network project on Hyperspectral Remote Sensing and Applications, Dr. Yogesh Simmhan, Assistant Professor, Department of Computational and Data Sciences, Indian Institute of Science, Bangalore and Dr. Soumya Kanti Ghosh, Chief Economic Adviser, State Bank of India. The panel noted that ‘we leave considerable digital footprints of the way we live and interact with the society and we are also witnessing increasing public availability of traditional information bases in digital form.’ The massive new information base at micro-level is waiting to be fully exploited for public policy. The multi-dimensional data in the presence of multiple actors in the financial and other domains makes enough interesting cases for Big Data analytics in both global and Indian context. The panel also discussed various ways to unleash the power of Big Data, which can help in unveiling the different phenomenon at ease – the privacy issues were also flagged in this regard. The discussion was concluded with the potential applications of Big Data in central banks.

Earlier, in his inaugural address Dr. Raghuram Rajan, Governor, (/en/web/rbi/-/speeches-interview/policy-and-evidence-1013), highlighted the need for using data and evidence to shape policy debates. Citing the examples of Reserve Bank’s recent efforts in moderating inflation and clean-up of bank balance sheets, he stated that the fragilities associated with high inflation accumulate, and eventually lead to crisis. Similarly, loan losses had a tendency to increase, get too big to ignore, too late to manage, and push the system into crisis. The current slowdown in the growth of non-food credit by public sector banks was largely because of stress in public sector banks, stemming from past mistakes in lending and the clean-up of their balance sheets, which is underway, needs to be taken to its logical conclusion. He said that the best way central banks can support growth over the medium term is by keeping inflation low and stable and successive governments, in their wisdom, have given the RBI a measure of independence in this regard. Such concerns also support the current government’s decision to enshrine its commitment to low inflation through a formal inflation target and the creation of a monetary policy committee.

Alpana Killawala
Principal Adviser

Press Release : 2016-2017/267