A1076
Title: A dynamic Bayesian network approach to interbank market
Authors: Wei Qian - University of Delaware (United States) [presenting]
Abstract: Motivated by the importance of understanding the nature of interconnections in interbank market networks and how the interconnections respond to changes in market conditions, a Bayesian dynamic interbank network model is proposed with probit link that simultaneously incorporates three underlying mechanisms for interbank trading: dynamic activity indices for overall market confidence, bank-specific latent variables for individual banks' fitness as borrowers or lenders, and pairwise covariates characterizing past trading relationships. Correspondingly, a computationally efficient Gibbs sampling algorithm is developed for sampling model parameters from the posterior distribution, which can be technically interesting by handling constrained parameters for model identifiability with latent components. By applying the developed Bayesian modeling technique to the e-MID interbank data analysis, estimation is obtained for empirically useful latent parameters that show not only satisfactory trading link forecasting performance but also gain in-depth insights into economic and pricing information for the interbank market. New model-based proxies of network topology change and relationship lending are specifically proposed to investigate their impact on relevant economic variables and the price of liquidity. The extension of the proposed Bayesian model to an alternative logit link is also discussed for modeling flexibility.
