A0235
Title: A Bayesian model for estimating spillover effects in unconventional monetary policy
Authors: Andrea Mercatanti - Sapienza University of Rome (Italy) [presenting]
Sharmistha Guha - Texas A&M University (United States)
Taneli Makinen - Bank of Italy (Italy)
Abstract: The focus is on a causal inference scenario where interference between units arises from equilibrium effects in markets. These are situations in which policy interventions, through modifications in supply and/or demand functions, influence equilibrium prices and consequently have an impact also on non-eligible units. The specific focus lies on equilibrium effects in bond markets resulting from an unconventional monetary policy implemented by the European Central Bank (ECB) between 2016 and 2018. This policy is known as the Corporate Sector Purchase Program, which aimed to further strengthen the pass-through of the Eurosystems asset purchases1 to the financing conditions of the real economy. The core idea of the proposal is to utilize data from markets that are likely unaffected by the policy under consideration, specifically the markets of Hong Kong and South Korea. This data can serve as additional controls, namely as observations that are not influenced by the policy either directly or indirectly. For inference, a Bayesian causal factor model is relied on where Bayesian shrinkage techniques are taken advantage of to simultaneously perform model selection and parameter estimation. In particular, the Bayesian Lasso shrinkage mechanism is employed after reparametrizing subject-specific factor loadings. Findings suggest that the Program had a statistically and economically significant negative impact on yield spreads for both the eligible and non-eligible bonds.
