Exploring the presence of Nonlinearities in the Peruvian Economy - Monetary Policy Implications

In this paper we identify different sources of nonlinearities in the Peruvian economy. For this purpose, we estimate five models: i) Linear Bayesian VAR (BVAR), ii) Time-Varying BVAR with Stochastic Volatility (SV), iii) Time Varying Mean BVAR with SV, iv) BVAR with SV and Volatility feedback, v) Threshold BVAR with SV and Volatility feedback. The results obtained allow us to conclude the following: i) The inclusion of data from the Covid-19 pandemic and later (2020 onwards) can be carried out safely even for a constant coefficients model, ii) SV (especially with volatility feedback) is enough to correct the downturn of the pandemic and other episodes of higher volatility. iii) The transmission mechanism of monetary policy is stable throughout the 2002-2024 episode, and is robust across different models, even for the pre-Inflation Targeting sample (1996-2001). iv) The estimated volatility for the models with feedback can be interpreted as an aggregate macroeconomic uncertainty index. This index reaches its highest value during the Covid-19 pandemic episode (2020-2021) and, to a lesser extent, during the International Financial Crisis (2008-2009). v) Shocks in volatility resemble those of a negative and persistent supply shock, where inflation rises, and the economic activity goes down. The latter triggers the response of the central bank through rising the policy interest.