The outcome of the US presidential election is the subject of this Metaculus question, which with over 330 interest votes at the time of question writing is the most "interesting" question on the front page of Metaculus by more than an order of magnitude. There also many related questions, which can be found here.
Several public forecasts for the election already exist, probably the most famous of which is this model by The Economist, which puts Trump's chance of winning significantly lower than the Metaculus community median.
Project 538's own forecast is widely anticipated, but has not yet been released, although an aggregated poll tracker is available. 538's own analysis of how well their model did in the 2018 midterms is available here.
This question asks:
When 538's presidential election forecast is released, will the chance of a Trump victory be higher than that given by the forecast from The Economist at that time?
In the case that multiple forecasts are released by 538, the model considering the most information is chosen for this question. For example, if, like in 2016, there is a "nowcast", a "polls-only forecast", and a "polls-plus forecast", the "polls-plus forecast" would be chosen. In the case that it is not clear for this description which model should be considered, 538's flagship model will be chosen using the algorithm described in a comment by @underyx.
If the models give the same probability when rounded to the nearest %, this question resolves as negative.
If 538 never releases an election forecast, or if The Economist withdraws theirs before 538's model is released, this question resolves ambiguously.