Basically, any state where the black population is more than 20% is a non-starter. This means that Alabama, Georgia and North Carolina automatically get crossed off the list. I know some people will want to argue about North Carolina, but the share of liberal whites won't be enough to overcome the black share of the electorate, which could be more than 45% of voters. Depending on how deep in the South you are, 20% of the population being black means that your Democratic primary voting bloc will be anywhere from 35% to 45% black.
In reality, I think any state where the black population is 15% or more could be thrown onto the same list, but I'm trying not to cross off too many states at once. The states that have black populations between 15% and 20% are Virginia, Arkansas and Tennessee. Arkansas is its own little ecosystem; obviously Clinton will win it even if the black population was only 10%. In these states, blacks will comprise anywhere from 25% to 35% of the Democratic primary electorate. I don't think NoVA is very friendly territory for Sanders, and so I expect his white vote here will be relatively weak. Tennessee (much like KY and AR) is really still such in a time-warp, and I expect whites will go quite resoundingly for Clinton along with blacks.
We're left with Oklahoma and Texas. I'm assuming Oklahoma is normal and has separate primaries for presidential and state/local. This will mean that a lot of those county-level Dixiecrats won't be voting in the presidential primary. Oklahoma is only about 8% black. Its presidential primary will likely be a little over 20% black; approximately two-thirds white. There'll be a few Latinos and Native Americans voting here as well, which could make a small difference if the margin were somewhat close.
Texas, however, is a different story in terms of diversity. Even though the black population is only about 12%, Latino participation in primaries is markedly lower. This will result in both white and black shares of the Democratic primary electorate being heavily inflated. It's completely possible that Texas' Democratic primary electorate will look comparable to, say, Tennessee's Democratic primary electorate when it comes to the share of the primary voters who are black (30-35% black); possibly even more. This is because Latinos will likely make up less than one-quarter of the primary bloc and whites here are about as Republican as they come.
If anyone is interested in doing some plug and play, then I highly recommend looking at my
"Who Comprised the Obama Coalition in the 2012 Election" map. This was done to determine racial breakdowns by state of Obama's voters in 2012. For primaries, my general rule of thumb is to reduce the Latino share of the electorate by at least one-third and re-distribute those percentages proportionately among the black and white segments of the Democratic electorate. That'll give you a general idea of what the primary electorate will likely resemble.
Poll
