We analyze the dynamics of the bank interest rates on the new short-term loans granted to non-financial corporations in seven countries of the euro area (France, Germany, Greece, Ireland, Italy, Portugal and Spain). Our specification is based on a multivariate diffusion model, involving factors and stochastic volatilities. In the application, we use a harmonized monthly database collected by the national central banks of the Eurosystem, over the period January 2003-November 2012. We estimate the model within a Bayesian framework, using Markov Chains Monte Carlo methods (MCMC). Unlike the results on spot rates in the empirical financial literature, we find that bank interest rates do not display evidence of mean reversion, and that the variance increases with the level of the bank rates only for a few countries. Moreover, we notice that the correlations between changes in the rates are not constant over the whole time period, and peak during the last months of 2008. Afterwards, they return more or less quickly to their previous level for some countries, while they remain lower for others. From this standpoint, the patterns within the euro area became more heterogeneous after the years 2008-2009.
Sanvi Avouyi-Dovi, Guillaume Horny and Patrick Sevestre
Classification JEL : E430; G210
Keywords : bank interest rates; diffusion model; stochastic volatility; Bayesian econometrics
Updated on: 05/03/2017 14:56