Agent-based Modeling to Benchmark Macroprudential Limits vs RW Add-ons

Abstract

Basel III has given rise to the active use of the macroprudential tools in the area of the banking regulation. The older and hence more frequently used one is the increase in the capital cost for the highly-risky loans (i.e., risk-weight add-ons). The more novel one is the quantitative limit without the direct capital cost implication. The regulators are limited in time to experiment when designing and introducing new regulatory tools. To respond to prudential regulators’ needs and to offer a regulatory toolbox for an appropriate experiment, we develop a banking system agent-based model (ABM). Due to experimenting with 70 ABM simulation scenarios, we are able to document the specific conditions that shape the efficiency of the macroprudential tools to limit credit and systemic risks, as well as to accumulate capital buffers to withstand prospective losses. First, we recommend considering the opportunity of new players’ entry. Second, we argue that the macroprudential tools are more efficient in normal times, rather than in the turbulent ones. Third, we recommend introducing banking-system-wide limits in addition to the bank-specific ones. We expect it to be of value for the regulators primarily in developing economies like Argentine, Thailand, as well as in part of developed ones like Ireland, Slovakia.