Sell-side analysts over-attribute a stock’s downside risk to macroeconomic forces. We document this Macro Bear Bias in 63,191 Morgan Stanley scenario-based valuation reports. Bear-case narratives assign 6.5 percentage points more weight to macro forces than base-case narratives by the same analyst, about the same firm, on the same date, while bull-case narratives show no such gap. The bear–base gap is largely insensitive to market conditions. Yet, conditional on the same market state, realized bear outcomes are no more market-driven than realized base outcomes. We explain the bias with a memory-based mechanism combining anchoring and representativeness. Analysts overgeneralize a remembered historical regularity—that severe declines are more macro-driven on average—to firm-specific downside scenarios under current market conditions, where the regularity need not hold. Consistent with this mechanism, the gap widened more after the Global Financial Crisis in the sectors most exposed to the crisis. The bias is consequential: the bear–base gap predicts pessimistic errors in subsequent base-case forecasts, and portfolios formed on a nonlinear bias-adjusted signal earn monthly CAPM alphas of up to 1.9%. Narratives analysts use to rationalize downside risk can distort numerical forecasts and asset prices.

Bayes Junior Asset Management and Asset Pricing Workshop, The Oxford-Man Institute (OMI) Machine Learning in Quantitative Finance Conference 2026, Research in Behavioral Finance Conference, FMA Annual Meeting, EUROFIDAI-ESSEC Paris December Finance Meeting