Since Doug Jones was elected to the Alabama Senate, changing its composition for 2018 to 51-49, I have spent some time analyzing whether Democrats could realistically flip control, now that only two seats will make the difference.
In pouring over data, one big trend stares out: midterm waves in the Senate group race results together - and it’s usually both against the party in the presidency and independent of the state and local factors.
Consider the following:
In 2006, under George Bush, Democrats gained 6 seats, from 44 to 50. They won in Montana, Virginia (then still reliably republican), Missouri, Ohio, Pennsylvania, and Rhode Island. Republicans did not win a single seat off the party out of power.
In 2010, under Barack Obama, Republicans gained six seats - in Illinois, Massachussets, Wisconsin, Pennsylvania, Indiana, and Arkansas - and Democrats won no hostile seats.
In 2014, also under Obama, Republicans gained nine seats - and Democrats again won no hostile seats. In fact, from 2008 to October 2016, Democrats net lost 14 seats with little reference toward state’s presidential voting leanings.
Add it up: in the last three midterm elections, across a 12 years, almost 100 individual elections, and a switch in presidential control, the party out of power has gained 21 seats from the other side. The party in power has gained 0. As you can see from the states list, the Republicans gained from blue seats and Democrats gained from red seats; there was essentially no correlation between state voting habits and winning - it was all about the national mood.
What does this mean for 2018?
If past is prologue, it means that the Republican advantage on the map means less than most analyses are suggesting. The generic ballot effect, which skewed heavily (+5 or more) to the party out of power in all three of those midterms, tends to swing close races late. So, for example, if Tammy Baldwin in Wisconsin is polling within 1-2 points of her opponent in Wisconsin and the margin of error but the generic ballot is +7 or 8, she should be considered a solid favorite to hold the seat. It’s not an exact parallel but Ralph Northam was polling only +3 at the end of the November Virginia governor’s race, then won by +9. The generic ballot swing was a clear leading factor.
So, part of this is also the simple mathematical combination of the generic ballot + incumbency. Incumbency is worth, on average, +7 over an open seat, and the current generic ballot is +10. So, any Democratic sitting Senator is comfortably handicapped +10-15 over a neutral race.
When you look at it from that lens, the map looks very different.