산업 내 준거점과 후회 심리: 시가총액 기반 후회 변수의 수익률 예측력

This study examines whether investor regret can explain cross-sectional variation in asset returns from a behavioral finance perspective. Traditional regret measures define regret primarily based on past returns, but investors in real markets may rely on multiple salient reference points such as market capitalization and trading activity. Prior studies show that regret based on the industry's highest return has predictive power; however, using a single reference point may not fully capture actual investor psychology. To address this limitation, we define new regret measures based on market capitalization (SIZEREG), trading volume (VOREG), and trading value (VOPREG), in addition to the conventional return-based regret (RETREG). Using Fama-MacBeth regressions for KOSPI and KOSDAQ stocks from 1986 to 2024, we find that all regret variables significantly predict future returns. SIZEREG shows the strongest explanatory power, suggesting that highly visible, large-cap stocks serve as important psychological reference points for investors. Multivariate regressions confirm that SIZEREG remains robust after controlling for firm characteristics and consistently outperforms RETREG. Double-sorting portfolio analyses further reveal that its predictive effect is stronger for small-cap, high-volatility, and low-price stocks. Robustness tests across alternative definitions and industry conditions support the stability of the results. Overall, the findings indicate that regret is shaped not only by past returns but also by attention-grabbing market features, offering meaningful implications for behaviorally informed investment strategies.