• Apr 08
  MS Teams

  Tuesday 8 April 2025, 10:30 AM

  Add to calendar

BOFIT Seminar - Tural Yusifzada (Innovation and Digital Development Agency) - Inflationmeter: A Non-Linear Analytical Tool for Predicting High and Low Inflation Regimes in BRICS Countries and Türkiye

Co-authors: Hasan Comert (Trinity Collage, Hartford) and Vugar Ahmadov (Central Bank of the Republic of Azerbaijan)

Abstract
This study introduces the “Inflationmeter,” a novel tool designed to predict forthcoming high and low inflation periods. In the first stage, we employ a Hidden Markov Model to establish high and low inflation thresholds for the BRICS countries (Brazil, Russia, India, China, South Africa) and Türkiye. In the second stage, we use a probit model to identify the non-linear, probabilistic relationships between inflation regimes and key drivers, such as unit labor costs, exchange rates, and global inflation. Our findings show that the “Inflationmeter” achieves over 90% accuracy in predicting inflation regimes across historical data, with particularly robust performance in the post-pandemic period. The proposed tool is considered a valuable analytical framework for central banks, especially in high- and low-inflation cases, which traditional linear or equilibrium-based models only partially predict.

Online BOFIT seminars are open to all researchers interested in the subjects covered. The seminar will be recorded and the recording will be shared only with registered participants. Those wishing to attend a seminar are kindly asked to register in advance, by filling in the Online Registration Form.

The registration for the seminar is open until 9:00 am the day of the seminar. You will receive a link to join the seminar by email at the latest one hour before the seminar is scheduled to begin.