Ethics & Public Policy Center

Courts Can’t Solve Our Gerrymandering Problem. Take It from a Former Gerrymanderer.

Published in The Washington Post on April 26, 2019


Courts are increasingly poised to become more heavily involved in determining what type of gerrymandering is constitutionally permissible. On Thursday, a three-judge court struck down 34 Michigan legislative and congressional districts as unconstitutional partisan gerrymanders. And later this term, the Supreme Court will decide on two partisan gerrymandering cases.

But take it from a former gerrymanderer: That’s almost impossible to do without injecting partisan bias into the decision itself.

I learned the secret craft of redistricting in college and got my first job on the basis of a map I drew for a 1983 California redistricting initiative. Stacking, packing, cracking — I’ve done it, seen it and secretly admired a good practitioner’s handiwork for decades.

Both parties pay tons of money for skilled practitioners because good gerrymanders generally work. This is clearly unfair, and I have always favored giving control over line drawing to commissions. But as much as we know unfair gerrymanders exist, it remains almost impossible to establish a clear, neutral standard by which a court can judge a particular map.

The biggest problem is establishing the correct partisan benchmark for the particular map. Voters have long cast different ballots for different types of offices. For example, in decades past, Republicans would win races for president or statewide office in the South, but voters would still vote for Democrats for Congress and the statelegislature. Comparing the number of seats a party wins in any year to any given statewide candidate’s benchmark could be accurate or inaccurate, depending on the candidate(s) and chamber(s) compared. Judges have no intrinsic way of discerning dross from gold.

Statewide seat totals are also influenced by questions of competitiveness that are not dependent on a map. A candidate needs to be able to raise a certain amount of money to be viable, especially if challenging an incumbent, regardless of party. If the party has a dry spell of poor hopefuls, the party could win fewer seats and receive fewer votes statewide than it should for reasons wholly independent from the district map.

Will judges be able to finely sift these effects from a map’s effects when considering a map’s constitutionality?

Personal popularity is also a strong factor, especially at the state legislative level. I can point to countless numbers of seats that should be held by a member of the opposing party but are not. The incumbent in these cases is considered so popular that he or she is not seriously challenged, and is sometimes not challenged at all. Yet any court case lumps these results in with every other seat’s and treats them the same. They are not, as any political pro would tell you, yet virtually no judge or judicial clerk has the political experience to identify those effects with any degree of sophistication.

Then you get the most problematic factor of all: shifting partisan loyalties. That’s what we are living through in the Age of Trump. Excellent examples in the dark arts of gerrymandering, drawn in 2011, now lie in tatters because millions of people have changed party loyalty. GOP gerrymanders assumed that wealthy suburbanites were Republicans, but those voters are now Democrats. Those “safe Republican” seats are now often held by Democrats or are dangerously close to switching. A map cannot be unconstitutional in 2012 and constitutional in 2018, but that’s what could happen under any standard that compares statewide vote to seat totals.

All of these uncertainties mean that judges will inevitably have to decide almost every set of lines on a case-by-case basis, dramatically increasing the perception that courts are simply tools of their parties. My experience is a perfect example of why this will be pervasive and corrosive.

In the early 1980s, Democrats predictably challenged a map I drew in California, arguing that it was an unconstitutional usurpation of the legislature’s prerogative to draw district lines. The California Supreme Court took up the case, at a time when the court consisted of six Democrats and one Republican. I was four years away from starting law school, but I did not need a law degree to tell me how the case would turn out. The 6-to-1 decision, on strict partisan lines, kept California’s voters from deciding whether they preferred my maps to Rep. Phillip Burton’s (D-Calif.) extremely well-crafted Democratic gerrymander. That work of “modern art” as Burton called it, added five seats to House Speaker Thomas P. “Tip” O’Neill Jr.’s (D-Mass.) majority for the remainder of the decade. Multiply this a hundred-fold in our hyperpartisan environment, and tell me it’s a good idea.

There is only one way to abolish partisan gerrymandering and keep the courts from getting involved in deciding whose partisan ox gets gored in each and every state: Adopt proportional representation for state legislatures and the U.S. House. If there are no district lines, there can be no gerrymandering. Set a low minimum threshold for getting seats, like Denmark’s 2 percent or Israel’s 3.25 percent, and there will almost never be a case where a party gets many more seats than they get votes. Switzerland has a bicameral system like ours, and has such a system for its House. Adopting that would fix our gerrymandering problem and fit our Constitution.

Supreme Court Justice Potter Stewart once famously wrote that he could not define hard-core pornography, but that he knew it when he saw it. So it is with gerrymandering, except judges don’t have the political experience to know what they are looking at. Take it from someone who does: The Supreme Court should either throw out the whole district system or stay out of the fray. If it doesn’t, Bush v. Gore-like decisions will become a recurring feature instead of an aberrant bug in our democracy.

Henry Olsen is a Washington Post columnist and a senior fellow at the Ethics and Public Policy Center.

Comments are closed.



RELATED PUBLICATIONS