Carol Shea-Porter was elected to represent New Hampshire in Congress in the Democratic wave of 2006. She won again in 2008, defeating Republican Jeb Bradley for the second time. In 2010, though, she lost to Frank Guinta, who replaced her. In 2012, they had a rematch, and this time Shea-Porter won. But last November, the two competed for a third time, and Guinta was re-elected.

If voting trends continue the way they have been, this might increasingly be what Congress looks like: Democrats getting elected in presidential years, Republicans replacing them two years later — and then the cycle starting over again.

Consider: The rumor of the day is that former North Carolina senator Kay Hagan (D) is considering a comeback in 2016. Hagan, you might recall, was unseated by Republican Thom Tillis in one of the closest Senate races last year -- during the election that brought Guinta back to the House and that swept Republicans to power in the Senate.

But why would Hagan think she'd win two years after she lost? Because of the electorate.

Since 1992, exit poll trends have looked like this for House voting. (Midterm elections are in lighter color.)

In no year did Republicans do better than Democrats in a presidential year. In only one year, in 2006, did Democrats do better than Republicans in midterms. (2006, we remind you, was when Shea-Porter first won.)

Consider: Earlier this month, former Wisconsin senator Russ Feingold announced that he planned a 2016 challenge to the man that beat him in 2010. 2010 was the great Tea Party wave. 2016 will almost certainly not be. There is also a 2010 rematch brewing in Pennsylvania, where Democratic former congressman Joe Sestak is challenging Sen. Pat Toomey (R).

Last spring, Sasha Issenberg detailed the two pools of voters in an article for The New Republic. There are the "reflex voters" -- older, whiter, wealthier and more Republican -- who come out every two years like clockwork. And there are the "unreliable voters" -- younger, etc., etc. -- who appear during presidential years and dramatically reshape the electorate. The 2014 election, Issenberg figured, would rely heavily on the reflex voters unless Democrats substantially increased turnout. Democrats did not substantially increase turnout.

If the pattern continues -- Democratic majority in presidential years, Republican majority in midterms -- there's no reason to think that the electoral pattern -- Democrat runs in 2016, Republican challenges in 2018 (or 2022) -- won't increasingly hold as well, at least in swing states and districts. We've only got a handful of candidates here, of course, so it's a bit early to predict that House seats will just change hands between two people cycle after cycle.

It happened in New Hampshire. And if Shea-Porter runs in 2016, she'd probably have a good shot at taking Guinta out once again. But in this case, that has little to do with the electorate.