with André Reslow
This paper introduces macroeconomic forecasters as political agents and suggests that they use their forecasts to influence voting outcomes. We develop a probabilistic voting model in which voters do not have complete information about the future states of the economy and have to rely on macroeconomic forecasters. The model predicts that it is optimal for forecasters with economic interest (stakes) and influence to publish biased forecasts prior to a referendum. We test our theory using high-frequency data at the forecaster level surrounding the Brexit referendum. The results show that forecasters with stakes and influence released much more pessimistic estimates for GDP growth in the following year than other forecasters. The actual GDP growth rate in 2017 shows that forecasters with stakes and influence were also more incorrect than other institutions and the propaganda bias explains up to 50 percent of their forecast error.