There's been quite a few stories about this of late. Today's New York Times has a lengthy piece about the Abrams-Kemp battle and Dan Balz' column in today's Washington Post assessed Abrams' chances against her Republican opponent.
So: could she do it? Sure, it's possible; the state has been trending in a direction favorable to the Democrats and she's a strong candidate in many ways. But we should be clear about just how difficult this is going to be.
To begin with the obvious, high black turnout will be essential to an Abrams victory. She may well get it. But that's highly likely to be enough to defeat Kemp.
There is a very simple reason for this. While the minority vote is large in Georgia, the white vote is much larger. It's unlikely to be under 60 percent of the vote and will probably be a bit higher.
Even in 2012, when Georgia black turnout was actually higher than white turnout (and way higher than white noncollege turnout), whites were still 62 percent of voters and blacks were just 32 percent.
Clinton in 2016, as many recent stories have noted, actually did better than Obama in Georgia, losing the state by just 5 points, compared to Obama's 8 point deficit. This improvement is entirely attributable to Clinton's improved performance among whites, both college and noncollege. Granted, her absolute support levels were still low among these groups, but her relative improvement was enough to make the state significantly closer.
But wasn't black turnout low in 2016 in Georgia? Yes, it did decline and slip below white turnout levels. But here's the thing. If Clinton had replicated Obama's 2012 black turnout levels in Georgia in 2016, she would have improved her margin by 2 points in the state. But, if she had replicated Obama's poorer 2012 white support in 2016, she would have done 6 points worse.
So the behavior of the white vote is key. As one might expect, the white college vote was where Clinton received the biggest boost in 2016 (a 14 point margin shift over Obama in 2012) and it is among this group that Abrams will be hoping to post big gains.
How big? In the Balz article, he quotes Democratic pollster John Anzalone as estimating that, if Abrams could get high black turnout, she could carry the state with only 30 percent of the white vote. That's the good news.
The bad news is that getting to 30 percent of the white vote is likely to be very difficult indeed. In 2016, Clinton got just 24 percent of the white vote. Assuming that Abrams does no better against Kemp among white noncollege voters than Clinton did against Trump, my back-of-the-envelope estimate is that Abrams might have to come close to splitting the white college vote evenly with Kemp (the size of the white college vote in Georgia is about two-thirds the size of the white noncollege vote). For comparison, Clinton got only about 35 percent of the white college vote.
So getting to 30 percent of the white vote ain't going to be that easy. But of course, it's not impossible and let's hope she's able to do it. A blue Georgia would be a beautiful thing.
The David Byler article linked to below provides more data that suggest the height of the hill Abrams will have to climb. Yes, this is an article from the Weekly Standard but Byler (a former associate of Sean Trende at RCP) is a smart, data-driven analyst whose stuff is always worth checking out.