Document Type : ORIGINAL ARTICLE

Authors

• Mohammad Mohsen sadr ¹
• Hossein Tebyaniyan ²

¹ Assistant Professor, Faculty of Educational Sciences and Psychology, Payam Noor University, Tehran, Iran.

² Assistant Professor, Department of Accounting, Payame Noor University, Tehran, Iran

Abstract

Monetary policy, as a key tool of macroeconomic management, plays a decisive role in financial stability and economic growth. As an important channel for transmitting monetary policy, the capital market reflects investors' expectations and reactions to changes in interest rates, liquidity, inflation, and exchange rates. In economies like Iran that face chronic inflation, currency shocks, and budget dependence on oil revenues, the capital market's response to monetary policies is asymmetric and nonlinear, and traditional linear models are unable to fully explain it. This study aims to fill this methodological gap by offering a combined approach of econometrics and artificial intelligence. In this framework, the structural vector autoregression (SVAR) model is used to identify causal and short-run relationships between monetary variables and capital market indicators, and then it is combined with a multilayer artificial neural network (ANN) model to discover nonlinear patterns. Monthly data for Iran during the period 1390 to 1403 were analyzed. The results showed that monetary expansionary shocks increase the growth of the total stock market index by about 4.7 percent in the short term, but have a negative effect through the inflation channel in the long term. In contrast, contractionary policies by increasing interest rates cause a 3.5 percent drop in the index in the short term, but reduce volatility by 22 percent in the long term. The combined SVAR-ANN model predicted the market reaction with 91 percent accuracy (R²=0.91).

Keywords

• Monetary policy
• capital market
• SVAR model
• artificial neural network
• machine learning

Main Subjects

• s