Saturday, September 9, 2017

Congressional retirements and 2018 forecasting

You may (or may not) have noticed the interesting news that a bunch of "moderate" Republican legislators have recently announced their retirement, like Charlie Dent.  In my actual, paying gig (the peer-reviewed writing process) I write about House elections, and for those of us in that realm, this is... maaaaybe big news.

OK, let's start with the fact that I put "moderate" in quote marks.  Susan Collins and Lisa Murkowski are moderates.  If you look up Dent's NOMINATE score (an ideology score based on voting patterns) from the last session of Congress, it was .242 on the -1 to +1 scale, with positive scores being conservative.  He is to the left of his party, but significantly to the right of Collins.  So, keep that in mind.

Anyway, what do these retirements mean?  Well...

Here's the deal.  Midterm elections favor the party out of the White House, barring something weird, like 9/11 boosting the president's popularity (creating the weirdness of 2002), or Congress impeaching the president over a blowjob (creating the weirdness of 1998).  Otherwise, the Democrats should be expected to gain seats in Congress.  The question is, how many?  That's... not easy to tell.  A couple of factors:

1)  The president's popularity matters.  Trump is very unpopular.  Historically so.  He loves bigness, so maybe that should comfort him!  Gallup has had him below 40% for a while, and it is hard to see him going above 40, without some external event.  Even with a war, it takes the other party's consent for a rally effect, as I have written before, so that presages badness for the GOP.

2)  The economy matters, but who the fuck knows where the economy will be?  If you know, please tell me so that I can plan my investments accordingly.  Also, I won't believe you if you tell me that you do know, so never mind.

3)  The map matters.  The House map intrinsically favors the GOP.  Now, I know you are thinking about the boogeyman of "gerrymandering!"  (Hard-g, people!).  Not quite.  Most of that is actually the simple fact that Democrats cluster in cities, making it so that you have to chop up cities in crazy ways in order to break down the Republican demographic advantage.  There are lots of factors that go into drawing district lines, and I'll be writing more about this as the Supreme Court case works its way through, but the House map does favor the GOP, mostly because Democrats just cluster in cities.  That makes it hard for the GOP to lose the House.  The Senate... that varies from year to year because only a third of the Senate is up in any given year.  2018 is the same set of seats that were up in 2012.  That's the same set of seats that were up in 2006.  Those were big years for the Democrats.  That's the map.  That means the Democrats are on defense, not offense.  It will be hard for the Democrats to pick up more seats in the Senate, given the map.

Anyway, that's a short list of the factors.  Now, incumbents are strategic.  Long ago, Gary Jacobson and Sam Kernell wrote a nice, little book, now out of print, called Strategy and Choice in Congressional Elections.  One of the observations in it was the idea that, when expecting a wave election, incumbents of one party might retire strategically rather than face a loss or a term in the minority, thereby opening up seats.  Incumbents very rarely lose, though.  This can sort of create a self-fulfilling prophesy.  There are problems with the argument if carried too far, and the model didn't really work to explain the biggest waves (1994, 2006, 2010), but the core observation is worth considering.

Incumbents win, depending on the year, 95% of the time or so in House elections.  The worst year for House incumbents?  2010.  Only 85% of House incumbents retained their seats.  Horrible odds, right?  There's no such thing as an anti-incumbent year.  Period.  Anyone who ever tells you that this year will be a bad year for incumbents doesn't understand the electoral process, no matter the year.  But, some years, incumbents of one party might get nervous.  Particularly the moderates.  They might retire strategically, thinking that it is better than losing, if they think it will be a bad year for their party.  Particularly if their president is... historically unpopular.

On the other hand, the map really does favor the GOP, so I wouldn't put too much stock in these retirements.  Could the Dems take the House, Senate or both?  Maybe, but right now, Trump's approval rating alone isn't necessarily enough.  Not with the map as it is.  Right now, we are just seeing a couple of actions that may or may not be "leading indicators," in economic terms.  Keep watching, but don't over-interpret them.

It is worth reading Dent's comments on why he retired, though...  He didn't talk about elections so much, although with Ileana Ros-Lehtinen and Dave Reichert also retiring, that's where my thoughts go.


  1. One of these days, I wonder if you want to work with me on this stalled, really old paper I had.
    I can't make the models of congressional elections that use a proxy for "number of vulnerable seats" work using a variable that better measures that (from a bunch of different cuts). So, the causal mechanism doesn't work when aggregated from the individual races (where it DOES work).

    1. I liked that paper. Happy to take a crack at it.