a quick primer on the data

When it comes to figuring out the political leanings of downballot districts, there really is no place like New Hampshire.  Literally, thanks to our 203 House districts!

The building block of being able to compare electoral districts to one another, and to understand what kind of headwinds or tailwinds candidates face in their districts, is called the Partisan Voter Index, or PVI.  

The concept of "PVI" was created by political journalist and pundit Charlie Cook in 1997, when he sought a way to compare congressional districts' partisan leanings to each other. He took the results of the prior two presidential elections nationally, averaged them, and then compared that average to the same process in every congressional district in the country.  For example, if a congressional district voted three points more for the Democratic presidential nominees in the last two elections than the country did as a whole, Cook would call that district "D+3".

That was revolutionary for its time, and it became commonly used language in political and journalistic circles.  But it was a pretty rudimentary way to understand the political leanings of a district, and it was only being used at the congressional level.  What about state legislative districts?

There are many methods currently used by the major political parties, operatives, and analysts to try to get at the same concept at the state level, but starting in about 2010, I informally started playing with my own way of trying to get at the most accurate version of this for downballot districts in New Hampshire.

Today, the PVI I use to rank districts and organize tiers of races (which can be found here) is based on a model that basically averages all statewide general elections results for the last three election cycles, takes out the non-major party results, and then recalculates the "D vs. R" percentages for each statewide general election.  Then, that average is compared to each House and Senate district in New Hampshire.  Those downballot districts are built by first doing that process for every precinct in the state, and then building districts from the ground up using the data from the precincts. It is critically important that all House and Senate districts weight the size of each town or ward in that district appropriately, or else the PVI of a House or Senate district will be inaccurate. 

It's a lot of work, but once the calculations are complete, it allows for almost endless levels of analysis and comparison. So, what are some of the most important things we can better understand because of this data?

Most importantly, it goes a long way towards understanding what kind of political environment our Democratic candidates enter at the outset of a campaign.

Over the past 12 years, I've found that when a House or Senate district is between D+6 and R+6, it is sufficiently "swingable" that it should be on our radar screen, at least to begin.  This doesn't mean that it will automatically be competitive - indeed, most races end up being won by whatever party has the partisan advantage (obviously?). How frequently?  During the last five election cycles (2012-2020):

Democrats have won:

- 100% of seats that were D+13 or greater

- 95.4% of seats that were D+6 to D+12

- 75.3% of seats that were D+0 to D+5

Republicans have won:

- 100% of seats that were R+13 or greater

- 94.9% of seats that were R+6 to R+12

- 61.4% of seats that were R+0 to R+5

In the handful of seats that were rated "Even", they have split exactly 50/50 between the two parties.

As you can see, Democrats have generally done better than Republicans in New Hampshire over the past decade at holding their own seats, and grabbing some of the other party's seats.

The problem has been - and continues to be - that because of the gerrimandering done by Republican majorities after the last redistricting in 2011, as well as the process completed this year, Republicans start with more seats that lean Republican than Democrats do:

In the 2012-2020 House district map, Republicans had 219 seats that leaned their way, while Democrats had 177.  (Four were even.)

In the new 2022-2030 maps, Republicans start with 228 seats leaning their way, while Democrats have 172. (None are dead even.)

So, from a Democratic perspective, this seems even less balanced now than it was a decade ago, right? Well, two things:

1) The reality is that the Democratic base is more geographically compact than the Republican base, and this hurts when it comes to drawing up districts. Democrats have more places where we win by enormous margins - think Portsmouth, Durham, Hanover, and Keene - but at the district level, you don't win extra seats in the legislature by winning a given seat by an extra-large margin!  This makes it easier to gerrimander us into hyper-Democratic districts, because those places tend to be cities and college towns. 

By contrast, Republicans have been the party of rural, exurban, and suburban voters, which tend to be a more "efficient" distribution of districts. Think of it this way: Democrats have more places where we will get 70% in New Hampshire, compared to the GOP - but Republicans have more places where they'll get 54%.

2) Compared to a decade ago, there are more districts that are in an emerging category of competitive suburban districts.  They remain, generally, Republican-leaning, but by less of a margin than a decade ago. This includes Bedford, Merrimack, Goffstown, Milford, Derry, Londonderry, and other (typically) southern-tier, sizable towns. 

This mirrors national political trends. President Biden's win margin in the key 2020 states came not from white voters without a college degree, or from non-white voters. It came from improved performance with suburban voters - a group that once was an essential part of Republican victories from Reagan right through George W. Bush. Indeed, much of the reason Republicans have been more successful with state-level races than federal races in New Hampshire in recent years is because of suburban voters' willingness to split their ballot (for example, Biden-Shaheen-Sununu in 2020).

The good news is that the trend lines suggest Democrats will continue to make inroads in these communities during the 2020s, which is critical - since New Hampshire has a higher percentage of higher-income, higher education voters than almost any state in America.  

One final benefit of this "NH PVI" is that it allows us to monitor trends within communities or regions over the years. Patterns clearly emerge within communities over time:  Places like Stratham and Hollis, which used to elect Republicans habitually, now actually lean Democratic, as their socioeconomic composition has changed (higher income, higher education). Blue-collar mill towns have also been trending - but towards the GOP, over the past 10+ years (Claremont, Berlin, Rochester). After the 2022 cycle, we'll take the updated election data, shift the window of data over two years to keep it current, and look for how each precinct's PVI has moved (if at all) over time.