Application of Instrumented Principal Component Analysis in Korean Stock Market

Abstract

This study applies the Instrumented Principal Component Analysis (IPCA) model to the Korean stock market to eva- luate its explanatory and predictive performance. Using monthly data from January 2002 to April 2023, we analyze non-financial firms listed on both the KOSPI and KOSDAQ markets. A total of 31 firm characteristics are incorpo- rated into the model, following methodologies established in prior literature. The performance of the IPCA model is assessed in comparison with observable factor models, including the Fama-French five-factor model with an addition- al momentum factor. The IPCA model consistently demonstrates superior explanatory power (total R²) and predictive accuracy (predictive R²), both in the full sample and across subsamples based on firm size, value, and beta. Out-of- sample analysis further confirms the robustness of the model, and investment strategies based on IPCA predictions yield statistically significant alpha even after adjusting for risk factors. These findings suggest that the application of IPCA framework is valid for asset pricing in emerging markets like Korea.