Monday, March 14, 2011

Updated hPVI by County

Over the next few weeks I'm going to be rolling out the new hPVIs, starting today with the counties and then moving on to the Senate and House districts. Because of redistricting the Senate and House hPVIs will be mostly irrelevant as soon as a new map is approved, but they will nonetheless provide some interesting information on partisan trends within the state.

For those of you who don't know what I'm talking about; hPVI is a bastardization of The Cook Political Reports PVI metric. PVI (which stands for partisan voting index) is simply the average margin of victory of a parties presidential candidate in a congressional district.

Here is the explanation from Cook's website:

A Partisan Voting Index score of D+2, for example, means that in the 2004 and 2008 presidential elections, that district performed an average of two points more Democratic than the nation did as a whole, while an R+4 means the district performed four points more Republican than the national average. If a district performed within half a point of the national average in either direction, we assign it a score of EVEN.

hPVI differs from PVI in that rather than using the last two presidential contests, I'm using the last presidential contest and the last Gubernatorial contest. And where PVI adjusts to the national average, hPVI adjusts to the state average.

Additionally, where PVI scores are formatted as D+2 and R+4, the hPVI values that follow don't have Ds and Rs in front of them. A positive value is a DFL lean, a negative vaule is a GOP lean. Feel free to reformat the numbers with Ds and Rs in front of them if you wish.

With that said, here are the new county hPVIs, DFL leaning counties on the left, GOP leaners on the right (Big Stone is even, but I put it in the DFL column simply for space reasons).

Image Hosted by

As was the case with the last iteration, their are far more GOP leaning counties in the state than DFL leaning counties. The rub is that the three largest counties in terms of population; Hennepin, Ramsey and St. Louis, are DFL stongholds. Before we get into that though, I want to look at the counties that experienced the largest partisan shifts between 2008 and 2010.

The chart that follows are the counties that experienced the biggest shifts in hPVI, those moving more to the DFL are on the left, those moving more to the GOP are on the right.

Image Hosted by

The partisan movement in the counties is interesting in that if you look at the GOP side, the majority of the movement is heavily GOP counties becoming more GOP, there are a few DFL leaning counties in there, but the list is primarily populated with already crimson red counties.

On the DFL side there are some heavily DFL counties that became more DFL, but there are also some GOP counties that moved to the left, like Steele and Roseau. What you can also see though, and this list is incomplete, is that there were far more counties that moved in the GOP's direction than those that moved in the DFL's direction.

This seems a bit strange considering that the data set that got replaced in the metric, the 2006 governors race, was a data set featuring a GOP victory, whereas the data set that replaced it, the 2010 governor's race, featured a DFL victory.

The explanation is actually quite simple; Hennepin and Ramsey counties got more democratic. When looking at just a counties hPVI, the thing that gets lost is population. Hennepin county at D+17 is far more important to the DFL than Carver county at R+28 is to the GOP.

To illustrate this point lets weight the hPVI value by population and see what happens. To do this I'll get the average share of the overall votes cast for each county and multiply that number by the hPVI.

So for instance; to get Hennepin county's weighted hPVI (or whPVI) we take the total votes cast in Hennepin county in 2008 and divide that by the total votes cast in the state. Do the same thing for 2010, average those two numbers together and we have the average vote share for Hennepin county. That number is than multiplied by the hPVI and we have whPVI.

Here's the whPVI list, with the DFL leaning counties on the left and the GOP counties on the right. At the bottom of each list a the total of that columns whPVIs.

Image Hosted by

This effectively illustrates the point about partisan density. The total whPVI of Hennepin, Ramsey and St. Louis counties is 7.11, the total whPVI of all 62 of the GOP counties is 7.52, not that much more than the total of the big three DFL counties.

If we apply the population weighting of a county to the partisan change we can see that the two point shifts of Hennepin and Ramsey counties have a much greater impact than do comparable shifts in the less populous GOP counties like Anoka and Wright.

Image Hosted by

The chart only contains the ten biggest movers on each side, but totals at the bottom are for all of the counties that shifted in that direction, not just those on the list.

If you were to take out Hennepin county the total shift would actually be slightly in favor of the GOP. But because of Hennepin county, and to a lesser degree Ramsey county, the overall shift was actually to the DFL's advantage.

This is the explanation for why Mark Dayton could win the Governor's race, yet the DFL could still manage to lose both chambers of the legislature. In a GOP wave election, where the majority of counties were moving in the GOP's direction, the Twin Cities was moving in the opposite direction.

The combination of Hennepin and Ramsey counties has averaged almost 32% of the share of the statewide vote in the last two elections. If the GOP wants to win statewide races they cannot afford to hemorrhage Twin Cities voters at this rate. There simply are not enough voters outside the Twin Cities to make up for the huge margins the DFL can rack up there.

Conversely, if the DFL wants to take back the state legislature they must do a better job in the suburbs and exurbs. To this point it's encouraging that Dakota and Olmsted counties moved in the DFL's direction even if that didn't translate to the legislative races.

Below is a chart that puts all this information together. It's similar to the Legislative hPVI chart I created previously. The column titled "hPVI" is the new hPVI number, the column titled "Old" is the previous hPVI value, the column titled "diff" is the difference between the two and "whPVI" is the population weighted hPVI.

The color coding is done so that for the "hPVI," "Old," and "whPVI" columns, the redder a cell is the more Republican it is, the bluer the more Democratic. For the "diff" column the redder or bluer it is the more it moved in that direction from the last version of hPVI to this one.

Image Hosted by

No comments:

Post a Comment