Classifying non-financial private corporate (NFPC) sector#–Issues and efforts

From manual profiling of a small number of large companies to construction of a frame for Industrial Activity Codes (IACs), from use of administrative databases (e.g. MGT-7) and other surveys (e.g. ASI) to exploration of Machine Learning for classification, the efforts to classify the NFPC sector have come a long way. Amidst the balancing concerns of comparability and representativeness, the reclassified estimates have been incorporated in 2017–18. While the impact of reclassification on relatively bigger segments like manufacturing has been minor, smaller segments like storage have been significantly impacted and share of other services (including the residual) has significantly reduced. Even though the classification based on IAC embedded in CIN has now been significantly reduced leading to activities being classified more in line with the actual present business, sources like MGT-7 also have to be treated with caution on multiple counts: lack of proper understanding of codes, skewed classification of vertically integrated businesses, erroneous classification as financial company based on property receipts in case of presently inactive business, etc. Probably, Machine Learning algorithms using industry-specific structural ratios coupled with administrative databases would help in improving the classifications further.