How do investors perceive the risks on individual stocks, and why do these beliefs matter? We recover analysts’ expectations of returns, variance, and skewness from bull, base, and bear forecasts in more than 180,000 reports from ten major brokerages. These expectations predict future volatility and asymmetry and support option strategies with Sharpe ratios near one. Consistent with asset-pricing theory, expected returns increase with perceived variance and decrease with perceived upside skewness. Scenario narratives help explain these higher-moment beliefs, with distinct information shaping variance and skewness. Applying machine-learning models trained on scenario narratives to 30 additional brokers’ reports produces text-based expectations that also predict future volatility, skewness, and option returns, extending the measurement of higher-moment beliefs beyond explicit price scenarios.