Understanding the impact of corruption on firm performance is crucial in designing policies that enhance the ease of doing business. However, the current firm-level literature on corruption is inadequate to generate concrete insights (Bahoo et al., 2021). One bottleneck is the unavailability of reliable data on corruption at the firm level. The second objective of this thesis attempts to solve this issue. The current most widely used corruption dataset at the firm level is the World Bank’s Enterprise Survey (WBES) which captures the experiences of about 145 countries with a standardized and comparable data collection technique. One limitation of the WBES data is that it is primarily qualitative and not collected annually to make it a panel dataset, hence limiting its usage and scope. Our study uses an innovative machine-learning methodology of the lasso to estimate bribery scores for organized Indian manufacturing firms and solve the data availability issue to some extent. While the current research focuses on India, the methodology can be easily extended and replicated for other datasets and economies. These estimated bribe scores are used for further analyses of objectives three and four of the thesis. In the third objective, we analyse the impact of corruption on firm performance, as measured by the intensity and efficiency of labour investments for the organized Indian manufacturing sector. We find a heterogeneous impact of corruption, empirically backed evidence for both the sand and grease hypothesis of corruption. These heterogenous impacts are based on varied quantiles of the performance variables, and the intensity of usage of other factors of production such as capital and energy. Corruption is often assumed to be a distortion that causes resource misallocation leading to reduced firm productivity. In the fourth objective, we analyse if corruption moderates the effect of misallocation on TFPG, i.e., if firms use corrupt methods to mitigate the negative impact of resource misallocation on their productivity. Our preliminary results support this argument for the case of Indian manufacturing firms.
Key Words: Corruption, Firm Performance, Productivity, Manufacturing Firms, India