The Dominance of Market Expectations over Systematic Risk on Stock Prices: Evidence from Fixed-Effects Panel Regression in Indonesia’s LQ45 Manufacturing Sector
DOI:
https://doi.org/10.31098/quant.4116Keywords:
Market risk, Investor expectations, Stock prices, Emerging markets, Panel data, Behavioral financeAbstract
This study investigates whether traditional market risk or forward-looking expectations dominate stock price formation in emerging markets, using evidence from Indonesia's LQ45 manufacturing sector. While previous studies generally examine risk and expectations separately, this research integrates both variables within a unified empirical framework. Using monthly panel data from five consistently listed firms during 2010-2023 (N = 840 observations), the study applies a fixed-effects panel regression model with Driscoll-Kraay standard errors to address heterogeneity, autocorrelation, and cross-sectional dependence. Market risk is proxied by beta and return volatility, whereas market expectations are measured using sentiment-based indicators. The findings reveal a significant asymmetry between the two determinants. Market risk is statistically insignificant (beta approximately 1.72; t approximately 1.32; p = 0.188), indicating that conventional risk measures fail to explain contemporaneous stock price variation after controlling for firm-specific effects. In contrast, market expectations show a negative and highly significant effect (beta approximately -19.13; t approximately -2.82; p = 0.0049), suggesting that pessimistic sentiment and expectation-driven factors play a dominant role in stock price formation. These results provide evidence that behavioral expectations outweigh traditional risk metrics in emerging markets. The study contributes to the asset-pricing literature by demonstrating that behavioral expectations serve as the primary transmission channel for valuation adjustments. Despite these contributions, the study is limited to a single sector and country context. Future research should extend the framework across industries, incorporate higher-frequency observations, and apply machine-learning-based sentiment measures to improve accuracy.

