The Jerome model published an updated prediction of the election result. It currently predicts a two-party vote share of 44.4% for Clinton, and 55.6% for Trump in Virginia. In comparison, on October 12, Clinton was predicted to collect only 44.3% of the vote.
In Virginia, the popular vote is often close. This is the reason why the state is commonly regarded as a battleground state, which makes it particularly interesting from a forecasting perspective.
Putting the results in context
As any other method, models are subject to bias. As a result, you should not focus too much on the results of a single econometric model. Instead of relying on results from single models, the evidence-based approach is to rely on combined models or, even better, a combined forecast that incorporates different methods and data.
The Jerome model compared with PollyVote's prediction
The current PollyVote predicts Trump to gain 45.9% of the two-party vote in Virginia. The results of the Jerome model for Trump are thus 9.7 percentage points above the combined PollyVote in Virginia.