Living With and Coping With Gerrymandering

Currently, the reapportionment wars are rampant across the nation. In some instances, the difficulties stem from political machinations. But some of these difficulties are inherent in the problem of trying to figure out a way to organize districts so that they reflect the relative power of the two parties within the state. In dealing with this issue, one common complaint is that the minority party within the state never receives its fair share of congressional seats due to the age-old practice of gerrymandering, which has been in place since the beginning of the Republic. There is some truth to that point, but the analytical difficulties are too often overlooked.

To see why, start with a perfect map drawn by a computer whose sole mandate is to equalize boundaries to avoid any and all political influence. Let us now assume that one of these two parties has 40 percent of the vote and the other has 60 percent. Assume further that this is a biggish state with 20 Congressional seats. Just how many Congressional seats would the minority party win under this setup? We do not know the answer until we know the distribution of the population by parties across several states. Start with the simple but incorrect assumption that the population of the two parties is uniform across all districts so that the local elections all come out the same way and the dominant party wins all the seats, which ironically is not surprising given that it also takes all the seats with 60 percent of the vote in all those states that use a winner-take-all-rule. And even in a state like Nebraska or Maine that have each district determine its separate districts, nothing changes. The dominant At this point, there is only one setting in which the minority party may capture any Congressional seats. It is to relax the homogeneity constraint and allow, as is often the case, the minority party to have better luck in some districts than in others. That should change the outcome for the non-gerrymandered district. But it does not guarantee that any district will move from 40 percent to over 50 percent of the vote. Indeed, the prevalence of incumbent victories suggests that such large swings are unlikely. So, the disproportionate distribution can easily lead to sharp differences. Here is a table from the Republican side that expresses their unhappiness with the current system in blue states.

The correct question to ask is what percentage of that disparity survives under the perfect regime mentioned above, given the current votes in the Electoral College. Ignore the tiny state with one or two Congressional seats.  There are also only modest options with midsize states like Massachusetts, with 11 votes, and Connecticut, with 6 votes, perhaps to bring a change in at most one or two seats. The places that look most susceptible to reform are the larger states, below the line. But parity is not obtainable under the current system in any of these states.

What can be done to improve the ratio? One obvious answer is to engage in a gerrymander with the sole purpose of creating those oddly shaped districts to propel the minority party. But before looking at that system, there is another less vulnerable solution. Districts with three Congressional seats and cumulative voting could mean that the minority party has a shot at one of those seats if its total exceeds one-third of the vote. Just that setup was used in Chicago when I first arrived there in 1972. The Republican candidate was Bernard Epton: “He was elected to the Illinois Legislature in 1968, thanks to the state’s unique cumulative voting system, which guaranteed the election of at least one Republican even in the most heavily Democratic districts, like Epton’s.” He quit the position when the single-districts were reestablished.

There is no apparent constitutional obstacle to cumulative voting in national elections where it functions to protect minority interests. There are obvious complexities about constituent services, but on balance, cumulative voting could do more to stop imbalance than anything else. And it is used in other institutional contexts for the same reason: to prevent the dominance of a single group on, say, a corporate board. But that prospect of diversification is exactly why the Democratic politicians killed it in Illinois.

To be sure, that part of the problem is with the inherent nature of step functions in voting. But so too is the gerrymander, which has its greatest power and works best in large states.

Guess which district (in Yellow) was cobbled to allow the Democrats in the state to improve the benefits that a dominant party gets by loading their opponents into a single district. All of this is to be expected because, as James Madison in Federalist No. 10 knew, “[e]nlightened statesmen will not always be at the helm.” His supposed cure was the extended republic to allow the best candidates to step forward, which was unduly optimistic given that clever opportunists can adjust their behavior to take advantage of the political realities. A more potent remedy is needed to counteract such a strong disease. One early success story occurred in 1960, in Gomillion v. Lightfoot, where a somewhat shocked Justice Felix Frankfurter struck down a 1957 Alabama Act invoking both the Equal Protection and Due Process Clauses (he didn’t care which). The act cut Tuskegee from a square into a “strangely irregular twenty-eight-sided figure,” solely “to remove from the city all save only four or five of its 400 Negro voters while not removing a single white voter or resident.” Obvious bad motive and an easy remedy: restore the original boundaries. 

The situation got murkier in Baker v. Carr (1962), where the Tennessee courts refused to order the state legislature to redraw district boundaries that had been unchanged since 1901. Justice Brennan found that this stonewalling violated the equal protection guarantee against the “dilution” of the votes of urban residents, who became a more beleaguered minority as the state expanded from 487,380 eligible voters in 1900 to 2,092,891 in 1960. He rejected the argument that Article IV, Section 4—“the United States shall guarantee to every state in this Union a Republican Form of Government”—should govern, because it offered no justiciable standard into which a court could sink its teeth. By finessing that clause, he did not have to deal with the common distinction between a republic and a democracy, where the former fears simple majorities and thus counters them with complex voting schemes, such as the direct election of senators in the original U.S. Constitution. By shifting gears to equal protection, Brennan claimed that he had a justiciable standard of one person, one vote for all contested elections, state or federal.

At the outset, one person, one vote was the only constraint. But once the race factor entered into the equation, the color-blind principle took a beating because of this (over-simplified) logic. If every district has the same ratio of white and black, say of 70-30, then each district will have a white representative, leading to an all-white House of Representatives, with black voters looking in from the outside.

Majority-minority districts were used to correct one problem, but they created a second. Justice Brennan’s one-person-one-vote standard became something of a misguided obsession, so tiny differences were ruthlessly suppressed, which in turn meant that numerically equal districts could only have virtual exactitude if they resorted to the same irregular shapes, only worse, that were shot down in Gomillion. In the federal system, the matter came to a head in Shaw v. Reno (1993), where the state of North Carolina had a 20-percent black population widely dispersed, and a 78-percent white population. Hence, to get a majority-minority district, one district had to look like a “Rorschach ink-blot” test, and a second was no wider than the I-85 for much of its length. The plan was rejected by a divided court, with Justice O’Connor concluding that it was “so irrational on its face” that it had to favor an illicit embrace of racial segregation. But at the remedial stage, she remanded the case for further consideration, leaving the applicable standards very much up for grabs.

In one way, the decision in Baker proved devastating for courts dealing with redistricting in cases where race was not the primary issue, such as traditional Red v. Blue contests. In these recurrent settings, the key insight was that as long as the population per district was kept equal, their shape could take any form, because race no longer operated as a partial restraint on district shape. At this point, the initial gap of Baker v. Carr leaped to the fore. Its fixation with exact population did little to stop some electoral abuse relative to a system where states could adopt, say, 5-percent deviations from parity to achieve a result that members of both districts might want—say, to allow residents of each of two neighboring areas to be in the same district as their fellow residents. Matters like this are relevant to constituent service, and in any event, do not open a serious door to abuse. But that numerical equality has done little to stop the gerrymandering that remains an ugly and pervasive fact in most states. One survey states that the Republicans in the South have greater luck than the Democrats in the North because they control more state legislatures and governorships. But in the minority of states where the Democrats are ascendant, they have not neglected to push their edge as the above table shows.

These errors accumulate to such an extent that a significant portion of the map becomes corrupted. The problem cannot be solved as in Gomillion by just undoing the new town map. Districts must be drawn, but how? The best plan is to pick a system that does not try to help any underrepresented groups, which only opens the door to more political shenanigans. So freed of all constitutional constraints, the best solution is that stated above: use compact districts drawn by a computer that seek to reduce the perimeter of any district relative to its area, and then to let the chips fall as they may, without even knowing or caring which party wins.

The dominant question is whether this abuse can be attacked by imposing some restraint on the shape and composition of districts. One notable attempt was Rucho v. Common Cause (2019), which sought to challenge a system that allowed the dominant party to engage in “packing” and “cracking,” as seen in Illinois. The former strategy involves drawing lines so that the members of the out-party are all concentrated in a single district with a major surplus of wasted votes. Cracking means that the dominant party divides the remaining districts, whereby it wins by far smaller majorities, so that every vote it gets counts toward putting together the winning majority. The plaintiffs limit the practice by measuring the “waste” votes given to the minority party. Not perfect, but it makes a dent in gerrymandering.

So if Baker v. Carr can impose the one-person-one-vote constraint, why not use the computer to stop grotesque districts? Chief Justice Roberts nixed the idea by mentioning in his opinion some possible ambiguities in this technical counterstrategy. But he paid no attention to the work on the topic led by Nicholas Stephanopoulos, one of the lawyers in the case for Common Cause, who explained how it could be best done. Instead, the Chief Justice made light work of the risks of nonintervention with partisan gerrymandering, which are all too evident today:

Vote dilution in the one-person, one-vote cases refers to the idea that each vote must carry equal weight. That requirement does not extend to political parties; it does not mean that each party must be influential in proportion to the number of its supporters.

Partisan gerrymandering claims present political questions beyond the reach of the federal courts. 

With that conclusion, he beat a hasty retreat from Baker v. Carr, where Justice Brennan did not bow to the global pessimism of Justice Frankfurter, who in Colegrove v. Green (1946) wrote that it was unwise for courts to step into the “political thicket.” Frankfurter was right that this was an objection, but wrong to turn it into a fatal one. We know now the full extent of gerrymandering under Baker v. Carr. The rising cascade of current abuse suggests that courts should be compelled to enter that thicket, but only with care.

There is yet a final irony: waiting in the wings is a census recount of eligible voters within each state, where the current evidence suggests that the problem is real. One exhaustive study from 2022 noted “States with estimated overcounts include Delaware, Hawaii, Massachusetts, Minnesota, New York, Ohio, Rhode Island, and Utah, while States with estimated undercounts include Arkansas, Florida, Illinois, Mississippi, Tennessee, and Texas.” There is a clear but not overwhelming Blue bias. Furthermore, major population shifts will further complicate matters, with large declines in votes in New York and California, against gains in Texas, with perhaps, according to the Wall Street Journal, a net swing of ten seats in 2030 to Red states, with major gains in Florida and Texas. Two-front wars often get bloody quickly, especially if the judicial referees are unarmed.

Richard A. Epstein is a senior research fellow at the Civitas Institute. He is also the inaugural Laurence A. Tisch Professor of Law at NYU School of Law, where he serves as a Director of the Classical Liberal Institute, which he helped found in 2013.