Liberalisation, concentration and diversification: Business Groups in India, 2000-2020- Data Resources
Description
We analyse the evolution of market structure in India between 2000 and 2020 using a rich dataset at high levels of disaggregation. We examine the extent to which business groups – notably family-owned groups – have sustained their dominant market positions in the Indian economy. We focus on two key dimensions. The first is the extent of concentration in markets and market shares by industry. The second concerns the dynamics and the extent to which business groups have focussed on consolidating their position in specific, narrow sectors or diversified by entering new sectors. We find that while market concentration has been falling, a bloc of high concentration sectors remains. Further, diversification has been actively pursued across sectors by most business groups. While this points to greater competition among business groups, the ratio of revenues to variable costs – a measure of the markup – has shifted upwards, particularly after 2013. The weight and persistence of these large business groups in the economy, as measured by the ratio of their revenues to GDP, has also increased. Finally, we discuss possible policy options. This dataset provides data for Figures and the STATA .do file for the replication of the replication of figures and tables. Due to restrictions on sharing the data by the data service provider "CMIE Prowessdx", the authors are not at liberty to share the dataset for the analysis.
Files
Steps to reproduce
The data set used for this analysis is the standalone financial statements data provided by CMIE Prowess. The data version is December 2023. However, the study only uses data from 2000-01 to 2019-20. Since the data usage agreement restricts the authors from sharing the dataset, this document outlines how the data set is prepared for the analysis. The CMIE Prowess variables from the standalone financial statements that are used for the analysis are: 1. Company name 2. Group owner code 3. Group owner name 4. Total income 5. Sales 6. Year of incorporation 7. NIC5 classification, required to generate the NIC3 classification. 8. Cost of: a. Raw materials b. Stores and spares c. Packaging and packing expenses d. Power, fuel and water charges e. Compensation to employees f. Outsourced manufacturing jobs g. Selling and distribution expenses h. Indirect taxes The dataset requires substantial cleaning for an analysis at the 2008 National Industrial Classification (NIC) two-digit level or higher. For example, 5025 firms are assigned to 20 non-existent or incorrect NIC-3 codes. This requires reclassifying these companies to their correct NIC-3 codes using the companies' financial statements, product-level revenue data for companies provided by Prowess which is used to identify the primary goods produced and an online source (Zauba Corp) which provides details for companies based on records from the Indian Ministry of Corporate Affairs. Our paper focusses primarily on the top 25 FBGs using total income of the business groups in 2019-20 as the ranking criteria. Total income is gross of indirect taxes, rebates and discounts and net of income capitalised and transferred to deferred revenues. The database reports total income at the subsidiary-level which is then aggregated up to the Family Business Group (FBG) level, our main unit of analysis. This is done in two steps. The first requires aggregating the subsidiary-level data for each business group at the NIC-3 level based on the group ID and name provided in the database. This step creates two types of business entities: Stand-alone Businesses (SBs) and Business Groups (BGs). The second involves aggregating BG branches to create the FBG entity. For this exercise, we consolidate BGs which are offshoots of the same broad family business and where the principals are close relatives and where connection between businesses exists. Specifically, this involves consolidation of branches for 9 FBGs . With this approach, we have a dataset where the total income of all firms is aggregated at the NIC-3 level and then consolidated. This gives three types of business entities: SBs, BGs and FBGs. All codes are provided in the STATA .do files are run using this data except for lines 309 to 332, which uses subsidiary-level data to identify the date of entry of a group-owned subsidiary in any NIC3 sector.