Business uncertainty in developing and emerging economies

We study business uncertainty in high- versus low-volatility environments by surveying over 31,000 managers across 41 countries. We elicit subjective probability distributions for future own-firm sales and measure firm-level uncertainty with their mean absolute deviations. Analogously, we measure realized volatility using absolute forecast errors. We establish two new facts. (1) Subjective uncertainty and realized volatility both decline with GDP per capita. (2) Managers underestimate volatility everywhere (they are overprecise), but more so in low-volatility rich countries. We build a heterogeneous-firm dynamic model and show that our facts imply larger TFP gaps between the US and developing/emerging economies. In the model, high volatility generates investment and growth opportunities in poor countries. But high uncertainty and low overprecision slow reallocation and pull down poor-country output. Quantitatively, the volatility effect dominates, so we infer 30 to 40% lower TFP in poor countries to reconcile their high volatility and low GDP per capita.