Although Democrats win 30% of the Utah House votes statewide but only 23% of the seats, we cannot conclude that partisan gerrymandering is to blame.
If you add up all the votes cast for Utah House candidates statewide in 2010, you’ll find that Democrats won 30% of the votes, yet they won only 23% of the seats. Democrats often attribute this disparity to gerrymandering.
In my previous post, I explained that single member districts (SMDs) can easily create these sorts of disparities. But even if SMDs are causing Democrats to win fewer seats (23%) than votes (30%), it’s also possible that a gerrymander is making the disparity even worse. So let’s take a look.
Rather than look at all 75 Utah House districts, let’s take a look at Utah’s two most populous counties: Salt Lake County and Utah County. To probe for any partisan gerrymanding in the 2001-2010 district maps, I’ll use precinct level data from these two counties. Unlike districts, precincts are not drawn for political reasons. Rather, they are merely administrative divisions. Salt Lake County administered the 2008 elections in 989 precincts; Utah County used 251.
There is not a single precinct in Utah County that comes close to partisan balance. The most left-leaning precinct in the county still gave an overwhelming 65% of its votes to John McCain in the 2008 presidential election.1
Suppose that we were to divide Utah County’s precincts into four legislative districts of equal population. If we took the most left-leaning precincts and grouped them together, they still would have given 74% of their votes to McCain in 2008. Meanwhile, if we took the most right-leaning precincts and grouped them together, they would have given 86% of their votes to the GOP.
There’s not a lot of daylight between 74% and 86%. We see that Utah County’s Democrats are too dispersed to exercise any sort of voting power. Although 29,567 (19%) Utah County voters supported Obama in 2008, they were too dispersed to win a single precinct.
These are exactly the circumstances under which SMDs can produce massive disparities in favor of the majority party. If Democrats are too dispersed in Utah County to win a single precinct, then Republicans will not need to gerrymander the map in order to deny Utah County Democrats representation in the legislature. SMDs alone will ensure that Republicans win every district, no matter how the maps are drawn.
Now let’s move from precincts to legislative districts. There are 12 Utah House districts that lie mostly within Utah County. In 2010, 20% of the votes in these 12 races went to Democratic candidates. Having won 20% of the votes, they might have expected to win 20% of the seats–that is, 2.4 seats. They didn’t. They won 0% of the seats.
Although Democrats won 20% of the votes but 0% of the seats, we cannot conclude that this was the result of a partisan gerrymander. It is simply the result of an unfortunate distribution of Democratic votes around the county. Utah County Democrats are not concentrated enough to win any seats in an SMD system. Even if Democrats were given complete authority to draw district lines in Utah County however they wish, there is no map that they could draw that would ensure a single district with anything close to a Democratic majority. If Democrats want to win any seats in Utah County, they should complain more about SMDs than about partisan gerrymandering.
Salt Lake County
So what about Salt Lake County? Because of its higher concentration of Democratic voters, Salt Lake County is much more prone to a partisan gerrymander than Utah County is.
Unlike Utah County, Salt Lake County does have strongly Democratic precincts. In 2008 its most Democratic precinct gave only 12% of its vote to McCain, with 88% of its vote to Obama.2
Suppose we were to divide Salt Lake County’s precincts into four districts of equal population. If we put the most left-leaning precincts into a single district containing 25% of the voting population, the most Democratic district would have given 69% of its votes to Obama in 2008. That represents a significant concentration of Democrats in Salt Lake County. Meanwhile, the most Republican district would have given 68% of its vote to McCain. We see that Salt Lake County is home to concentrated Democratic populations and also concentrated Republican populations.
Of course, we are completely ignoring contiguity. We cannot just put a bunch of precincts into a single district unless they are physically contiguous (i.e. adjacent). Still, this illustration shows that there is sufficient Democratic concentration within Salt Lake County to win elections, even in an SMD system. That’s a major difference from Utah County.
Let’s move from precincts to legislative districts. There are 31 Utah House districts that lie mostly within Salt Lake County. In 2010, 45% of the votes in these 31 races went to Democratic candidates. Meanwhile, 16 of 31 races (52%) were won by Democrats. In Salt Lake County, Democrats are actually winning slightly more seats than votes. That’s exactly the opposite of what happens in the rest of the state.
It’s safe to conclude that Salt Lake County’s districts were not drawn in a way that gives obvious advantage to Republicans. If Republicans had gerrymandered Salt Lake County, then Democrats would not be winning more seats than votes. Instead, the SMD system is actually helping Democrats in Salt Lake County, just as it helps Republicans in the rest of the state. (It also helps Democrats that their districts are currently underpopulated; see details here.)
What do we learn?
In Utah County, Democrats won 20% of the votes but 0% of the seats in 12 Utah House races held in 2010. In Salt Lake County, Democrats won 45% of the votes but 52% of the seats in 31 races.
In an SMD system, the geographic distribution of voters matters a lot. Because Democrats are dispersed in Utah County, there is no district map imaginable that would allow a Democrat to coast to victory. Democrats’ poor performance in Utah County is caused by SMDs, not gerrymandering.
Meanwhile, Democrats are concentrated in Salt Lake County. A Republican gerrymander would break up these concentrated Democratic populations, causing them to win fewer seats than they deserve. Clearly, there is not a Republican gerrymander hurting them, since Democrats there are winning even more seats than votes.
Let’s extrapolate from these two counties to the entire state. Although Democrats win 30% of the Utah House votes statewide but only 23% of the seats, we cannot conclude that partisan gerrymandering is to blame. It looks from these data that the geographic distribution of Democratic voters has a lot more to do with this outcome than the district map does.
This is part of a series of posts about redistricting in Utah. For an overview, read the introductory post. My talented research assistant, Robert Richards, contributed heavily to this series.
The title is misleading. It doesn’t discuss gerrymandering of 2001, just voting patterns of 2008 and 2010. It also talks about presidential elections and state legislative seats, not congressional districts which is what the 2001 maps were most criticized for. It would be interesting to have precinct data analysis of the congressional districts before and after.
Will, read the title again. The title asks whether legislative districts were gerrymandered, not whether Congressional districts were gerrymandered. I’m using voting patterns in 2008 and 2010 to look for evidence of gerrymandered legislative districts.
As it happens, I did write an analysis of the Congressional districts, but that’s in a different post.
Ah, I see things clearer in the light of day. I’d actually read all of the articles before.