By my calc, in the past four years of Census data, non-college whites - Trump& #39;s base - have declined from 47.0% to 44.5% (-2.5%) of the nation& #39;s voting age citizens, including:

AZ: 44.9% to 42.0% (-2.9%)
FL: 45.1% to 41.9% (-3.2%)
TX: 35.1% to 32.3% (-2.7%)

By my estimate, if turnout levels & partisan preferences among 1) non-college whites 2) college whites and 3) non-whites were to remain *constant* from 2016, Biden would flip:

Florida (29 EVs)
Michigan (16)
Pennsylvania (20)
Wisconsin (10)

...based on demographic change alone.

Moreover, in AZ, demographic change alone (again, turnout/partisan preference remaining constant) might cut Trump& #39;s 2016 margin by more than half, from 3.5% to 1.6%.

In GA, demographic change alone might cut Trump& #39;s 2016 margin from 5.1% to 3.5%. In NC, from 3.7% to 2.4%.

Of course, there are a lot of caveats here. For example, there& #39;s some evidence Trump/Rs have actually recovered ground w/ Latinos in FL since 2016, hence R wins there in 2018.

And overall, demog change is still only a small component of Trump& #39;s weaker standing relative to 2016.