The aim of this paper is to investigate heterogeneity in macroeconomic news forecasts using disaggregate data of monthly expectation surveys conducted by Bloomberg on macroeconomic indicators from January 1999 to February 2013. We find three major results. First, we show that macroeconomic indicator forecasters are mostly heterogeneous and their expectations are found to violate the rational expectation hypothesis. Second, the use of the expectation mixed model –combining extrapolative, regressive and adaptive components– reveals a large dominance of the chartist profile among forecasters with a systematical persistence over time despite all the structural breaks determined endogenously by the Bai-Perron estimation method. Third, we find that forecasters whose forecasting models combine at least two or three anticipatory components (extrapolative, and regressive or/and adaptive) and display high temporal flexibility, thus adapting to different structural breaks, are those which provide the most accurate forecasts.