Investor sentiment, accounting information and equity valuation in Latin American emerging markets
Márcio André Veras Machado, Lineker Costa PassosPurpose
This study aims to examine whether investor sentiment affects two key valuation inputs, expected earnings growth rate and discount rate, and whether it moderates the relationship between accounting information and firm value.
Design/methodology/approach
The study analyzes firms from Brazil, Chile, Mexico and Peru between 2004 and 2023. Expected earnings growth relies on analysts’ earnings per share forecasts, the discount rate on the implied cost of capital and investor sentiment is measured at the firm level.
Findings
The findings show that investor sentiment meaningfully alters equity valuations by influencing expected earnings growth and the discount rate, generating deviations from intrinsic value. Sentiment also changes how accounting information is incorporated into prices. Its economic impact is sizable, exceeding that of traditional risk factors such as beta.
Research limitations/implications
The sentiment measure may not fully capture investors’ beliefs, and results may not generalize to markets with stronger institutions. Still, the findings imply that policymakers should reduce informational frictions that amplify sentiment-driven mispricing, while managers should monitor sentiment as an additional factor shaping market reactions to fundamentals.
Originality/value
By integrating behavioral finance predictions with the residual-income framework, the paper shows that sentiment-driven expectations can rival traditional risk-based determinants in shaping valuation inputs. This provides new evidence for the Latin American context, where political instability, regulatory heterogeneity, and persistent information asymmetries heighten investors’ reliance on heuristics. In such environments, sentiment-induced shifts in expectations more readily propagate into growth forecasts and discount rates, helping explain the stronger firm-level sentiment effects observed in the sample.