Not Just Sentimental Value! The True Value of Sentiment in Financial Forecasting
摘要
This study investigates the sentiment expressed in minutes of the Monetary Policy Council (MPC) meetings regarding the nominal interest rates setting, examines its determinants, and usefulness in forecasting. The sentiment is quantified with the VADER analysis tool. Its key determinants include EURPLN exchange rate, inflation rate, and WIG20 stock index price. Forecasts obtained with machine learning indicate that sentiment scores reliably predict MPC decisions on interest rate changes in Poland.