Connected Lending and Aggregate Productivity

Abstract

(ProQuest: ... denotes formulae omitted.)IntroductionConnected lending occurs when financial intermediaries grant loans to some firms based on their special connections but not on firm characteristics. Bualek (2000) argues that the founders of most Thai commercial banks established their banks in order to channel loans to their own non-bank businesses. When connected lending arises, resources are not allocated to their best uses. Restuccia and Rogerson (2008) show that resource misallocation can considerably decrease aggregate output and aggregate productivity. Since connected lending can cause resource misallocation, connected lending can severely and adversely affect aggregate productivity.There is actually some evidence of the relationship between connected lending and aggregate productivity in the cross-country data. The World Bank Business Environment Survey (WBES) which was conducted in 2000 has a survey question related to the issue of connected lending. In the survey, firms are asked how problematic the need for special connections with banks or financial institutions is for the operation and growth of their businesses. The variable takes a value between one and four. One means no obstacle, two means minor obstacle, three means moderate obstacle, and four means major obstacle. So, the higher the value means the more of an obstacle the issue is and thereby means the more prevalence of connected lending. The need for special connections variable taken from the WBES, therefore, can be used as a measure of connected lending.Figure 1 plots real gross domestic product (GDP) per worker obtained from Caselli (2005) against the country average of the need of special connections, which is a measure of connected lending. As can be seen from the figure, the issue of connected lending is negatively associated with aggregate output.Figure 2 plots total factor productivity (TFP) obtained from Caselli (2005) against the country average of the need of special connections, which is a measure of connected lending. As can be seen from the figure, the issue of connected lending is negatively associated with aggregate productivity.This paper uses a model economy embedding connected lending to quantitatively evaluate the impact of connected lending on aggregate productivity. The model incorporates entrepreneurship, financial frictions, and connected lending into an otherwise standard neoclassical model. In the model, individuals choose whether to become an entrepreneur operating an individual specific technology or to become a worker supplying labor for a wage. This occupation choice allows for endogenous entry and exit decisions in production, which are important sources of resource misallocation. Individuals differ in their productivity, wealth, and special connection status. Individuals' wealth is determined endogenously by their forward-looking saving decisions. The financial frictions are modeled as collateral constraints. Entrepreneurs with special connections face a more relaxed collateral constraint compared to those without special connections. These collateral constraints and the differentiation in these constraints among entrepreneurs limit efficient reallocation of resources across entrepreneurs. The model is then calibrated using data on standard macroeconomic aggregates, establishment size distribution, establishment dynamics, the concentration of income in the population, and firms' external financing.Economies in this study differ from one another in their degrees of connected lending. There are several ways to model the differentiation in collateral constraints between entrepreneurs with and without special connections. One way is to loosen the collateral constraint faced by entrepreneurs with special connections but tighten the collateral constraint faced by entrepreneurs without special connections. This way of modeling allows the external finance to GDP ratio to be kept roughly constant in order to isolate the effect of connected lending. …