UTILIZATION OF MACHINE LEARNING FOR EMPIRICAL ASSET PRICING IN EMERGING MARKETS
No Thumbnail Available
Date
2024-12-12
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Nazarbayev University Graduate School of Business
Abstract
We perform Principal Component Regression (PCR) analysis to predict cross-sectional stock returns in emerging market economies. As a benchmark comparison, we employ OLS models and demonstrate predictive power of the machine learning based PCR model. We utilize 64 firm characteristics to determine the most significant predictors for the emerging market countries as well as for individual countries. The results, demonstrate predictive power of the PCR model over the linear regression model, showing consistent results in both the country-specific analysis and in the overall analysis of the emerging market. The most important set of predictors throughout the analysis proved to be book-to-market, sales-to-price, leverage (lev), cash flow-to-price (cfp), dividends (dy), and gross profitability (gma).
Description
Keywords
Principal component regression, Emerging markets, Machine learning, Stock returns, Type of access: Open
Citation
Adilov, D. (2024). UTILIZATION OF MACHINE LEARNING FOR EMPIRICAL ASSET PRICING IN EMERGING MARKETS. Nazarbayev University Graduate School of Business