The increasing polarization in political issues over the world calls for a better understanding of how people form and express their beliefs (Canen et al, 2020). I ask whether peoples’ forecasts about important economic indicators (e.g., inflation) are affected by their political preferences (i.e., biased). I conduct a pre-registered online experiment in a high uncertainty political context given by the 2023 presidential elections in Argentina. The experiment exogenously manipulates the incentives to report accurate forecasts about economic indicators and information about the current levels of the same indicators. I find that incentivising responses reduces the gap between subjects' forecasts regarding the performance of different presidential candidates and explore what drives the results.