COVID-19 and Local Market Power in Credit Markets

This paper investigates how COVID-19 affected the local market power of Brazilian credit markets. We first propose a novel methodology to estimate bank market power at the local level. We design a data-intensive method for computing a local Lerner index by developing heuristics to allocate national-level bank inputs, products, and costs to each branch locality using data from many sources. We then explore the exogenous variation in COVID-19 prevalence across Brazilian localities to analyze how the pandemic influenced local market power through the effective price and marginal cost channels. Despite reducing the economic activity substantially in more affected localities, COVID-19 did not significantly impact the effective price channel: bank branches offset the decrease in credit income by reducing credit concessions. However, bank branches more affected by COVID-19 experienced increased marginal costs as they could not rapidly adjust their cost factors in response to the decrease in credit concessions. Consequently, COVID-19 reduced banks’ local market power via the marginal cost channel. However, banks that spent more in IT before the COVID-19 outbreak suffered less replacing more easily local borrowers with remote ones. We then design a bank-specific measure of exposure to COVID-19 to examine how the pandemic affected different banks within the same locality. Banks more exposed to COVID-19 increased their local market power mainly via the effective price channel, which operated through a negative supply shock and not increased credit income. The paper provides new insights as to how crises can affect local market power in non-trivial ways.