Enhancing macroeconomic forecasts with uncertainty-informed intervals

We propose a methodology for constructing confidence intervals for macroeconomic forecasts that directly incorporate quantitative measures of uncertainty – such as survey-based indicators, stock market volatility, and policy uncertainty. By allowing the width of confidence intervals to vary systematically with prevailing uncertainty conditions, this approach yields more informative and context-sensitive intervals than traditional, static methods relying solely on past forecast errors. An empirical application using Austrian data demonstrates that uncertainty measures significantly explain the variation in forecast errors, underscoring the value of integrating these indicators for improved communication and analytical robustness of economic projections.