Does gerrymandering matter? Not as much as you might think. You're sure to be wrong if you take at face value the rhetoric of liberals who seem to place most of the blame for Republican majorities in the House of Representatives on partisan map-making.
I have argued repeatedly — in December 2017, September 2017, July 2015, October 2014, September 2014, January 2014, and February 2013, that redistricting is less important in securing Republican congressional and legislative majorities than demographic clustering — the fact that Democratic voters are increasingly concentrated in black, Hispanic, gentry liberal and university areas.
That's because Democrats’ huge majorities in districts dominated by such voters do nothing to elect Democrats in the remaining districts. A party with clustered constituencies is inevitably disadvantaged by a system of equal-population legislative districts. That conclusion is confirmed by the research of political scientists Jowei Chen (University of Michigan) and Jonathan Rodden (Stanford), as reported in the New York Times in 2014, and it was confirmed once again last week by the work of David Wasserman and three colleagues at FiveThirtyEight in their Atlas of Redistricting.
Chen and Rodden used a computer algorithm to draw thousands of districting plans based on 2010 Census data and used the results of the 2008 presidential election (the last available to redistricters operating in 2011 and 2012 with the 2010 Census data) to estimate how many seats each party would win. They found that random drawing of districts would give Democrats “only a small handful of seats” — probably just single digits — more than they won in the actual districts in effect in 2012.
David Wasserman and his colleagues have performed a similar exercise. They have programmed computers to perform several thousand congressional redistrictings for all 43 states with multiple congressional districts (seven states have only one district) under seven different assumptions — Democratic gerrymander, proportionally partisan (the standard sought by Wisconsin Democrats in the redistricting case currently before the Supreme Court), majority-minority (the standard required by the long prevailing interpretation of the Voting Rights Act), highly competitive, compact (borders), compact (algorithmic), current, and Republican gerrymander. Then they estimated likely partisan outcomes based on recent election results.
Only one of these seven methods — Democratic gerrymander — produces more “usually Democratic” than “usually Republican” districts. Even proportionally partisan plans, which Democrats are hoping Justice Kennedy and the four Democratic-appointed justices will require, produces more usually Republican than usually Democratic districts. The current plans, by the way, produce 195 districts they classify as "usually Republican" and only 168 as "usually Democratic."
For more details, consult the Atlas of Redistricting. And take a look at Wasserman’s FiveThirtyEight article titled “Hating Gerrymandering Is Easy. Fixing It Is Harder.” In the process, he confirms points I have been making for more than five years and which others, including Wasserman himself, have increasingly been making.
Why do Republicans tend to have an advantage in redistricting, and why do partisan Democratic districting plans tend to produce less compact and aesthetically pleasing district boundaries? “First,” Wasserman writes, “more than in past decades, Democratic voters are inefficiently clustered in big cities and college towns.” Bill Clinton’s 49 percent 1996 coalition was spread more evenly around the country than Barack Obama’s 51 percent 2012 coalition.
“Second, the Voting Rights Act limits the extent to which Democrats can spread their voters across many districts, because it provides safeguards against diluting majority-minority districts.” The courts have been finagling about the rules and chipping away at the long prevailing interpretation of the Voting Rights Act that required maximizing the number of majority-minority districts; but this is still a handicap for Democrats.
Wasserman and his colleagues, FiveThirtyEight journalists Aaron Bycoffe, Ella Koeze and Julia Wolfe, have produced splendid work. It should be required reading for anyone interested in understanding redistricting, and for those — including Justice Anthony Kennedy — who hope there is or think there must be some easy answer, some computer algorithm or legal formula, which will guarantee fair, non-partisan redistricting. There isn’t.
The best solution, and one I have long advocated, is strict application of the equal-population standard adopted by the Supreme Court in 1964, which limits more stringently than most observers think the ability of redistricters to benefit their party or faction. This also has the advantage of being easily, ministerially (a legal term) enforced by the courts in a non-partisan manner.