Some Weaknesses in Current Arguments Against Gerrymandering

In anticipation of redistricting following the 2020 census, in light of several pending court challenges, and possibly because Republicans are seen as the current beneficiaries, there has been much comment in the news media about political gerrymandering. Although I agree that the ideal means of drawing district boundaries would not consider the partisan make-up of the district, the current criticisms of gerrymandering are inadequate because they do not propose sufficient standards as to how districts should be drawn and seek a judicial remedy without an adequate textual basis in, or manageable standards drawn from, the Constitution.

To start with basics: Where a legislature consists of representatives of particular geographic constituencies that do not have any significance apart from the selection of legislators, it is necessary for someone to draw the boundaries that will define those constituencies, typically called “districts” in American parlance, hence the term “redistricting.” The term “gerrymander” in relation to the drawing of districts to maximize the electoral success of one party, comes from criticism of Massachusetts state senate districts established by legislation signed by then-governor Elbridge Gerry (a signer of the Declaration of Independence, delegate to the constitutional convention who voted against the Constitution, and vice president of the United States).

In our two-party system, expected voters for each party are not equally distributed on a geographical basis. Some areas will have more Republicans than Democrats, and vice-versa. A perfectly gerrymandered district layout would concentrate the expected voters for the party not controlling the redistricting process into as few districts as possible where its candidates would be expected to win by large majorities. Expected voters for the controlling party are, on the other hand, put into districts where they have much smaller, but still safe, majorities. Modern statistical methods allow for very gerrymandering to be very effective. This, the editorials tell us, is bad because electoral results should mirror the votes case for each party. (I am not convinced that a vote for a candidate running under the “Party A” label is necessarily the same as a vote for Party A to have a legislative majority. Voters can choose individuals despite their party labels.)

As a very simple example, imagine a state with 500,000 voters equally divided between supporters of Party A and Party B (250,000 each) and the need to create five districts, each with 100,000 voters. (In this example I am ignoring the distinction between voters and non-voter residents.) If Party A is in control of the redistricting process, it would seek to pack as many Party B supporters as possible into one district (let’s say that the geographic distribution allows for a district of 90,000 expected Party B voters and 10,000 Party A voters). That leaves 240,000 expected Party A voters and 160,000 expected Party B voters. If Party A can allocate each group equally into the four remaining districts, it is to be expected that Party A will safely obtain four of the five districts in every election, despite equal support in the state as a whole.

Redistricting does not affect the United States Senate because the districts from which Senators are chosen are whole states whose boundaries are not subject to change for the purpose of choosing Senators. Redistricting is also inapplicable to smaller legislatures whose members are chosen on an “at large” basis.

While it is theoretically possible to base legislative representation on otherwise existing geographical units, the practice has been found to be unconstitutional because it results in districts that are too far divergent in population, a consideration that trumps other concerns. (For example, as I mentioned elsewhere, in the New Jersey Senate, each county was formerly represented by one senator, until the rejection of the so-called “federal analogy” by Reynolds v. Sims, 377 U.S. 533, 84 S. Ct. 1362, 12 L. Ed. 2d 506 (1964), eventually resulted in the current structure, in place since the 1973 election: 40 legislative districts, with substantially equal population, each electing one senator and two members of the general assembly.) The composition of the United States Senate is specifically set forth in the Constitution and therefore it is probably the only elected legislative body in the United States that is not subject to a population equality standard for its electoral districts.

It is a basic fact of human existence that people will tend to act in what they perceive to be their own self-interest. Politicians are hardly immune from this tendency. Therefore, given the opportunity to design favorable districts, the party controlling the redistricting process will naturally seek to create them. Politicians will avoid drawing new districts that pit incumbents against each other, but this is also a factor likely to result in oddly shaped districts

The need for substantial equality of population requiring the drawing of district lines and generally overriding other limits on how the districts are delineated (such as respect for the boundaries of political subdivisions) makes it difficult to articulate standards that might make gerrymandering less effective. Most of the recent editorials bemoaning political gerrymandering and applauding or hoping for judicial intervention fail to articulate standards by which districts are to be created or structural methods for their creation.

What are the important affirmative standards that should inform the drawing of district lines? If respect for boundaries of political subdivisions or areas defined by natural boundaries is important, do we not have the technological tools to judge redistricting plans (assuming that the proposed districts meet a reasonably standard for population parity) by the degree to which they show this respect and use that as the basis for choosing the best?

If, on the other hand, the important thing is that the electoral fortunes of each party reflect the votes cast for its candidates, then the time has come to abandon election from districts and to adopt a proportional voting system, such as a party list or multi-member constituencies with a transferrable vote as in the Republic of Ireland.

While there might be general agreement that gerrymandering is not the best way to draw legislative districts, the present debate as reflected in editorial pages has not identified any constitutional basis to prohibit it. Courts may override legislative judgments only as a matter of constitutional imperative, not because they believe some other way would be better , or even more democratic.

This lack of an affirmative guide for the drawing of districts is part of the reasons that courts are ill-suited to devise a remedy or to draw districts themselves. Absent an ascertainable standard drawn from the text of the Constitution, such plans, especially where judges are elected, are likely to be no less political than those emanating from the other branches.

Thus, the remedy for political gerrymandering is not simply to ban or limit the practice, but to articulate and adopt a comprehensive and coherent set of principles for how districts (if there are to be districts) should be drawn. Only then can there be manageable objective standards by which to evaluate any particular redistricting scheme.

A Problem of Bicameralism

When the United States fought its war for independence from Great Britain, the various colonies sent delegates to an assembly known to history as the Continental Congress. This loose organization was formalized by the adoption of the Articles of Confederation, which called its legislative body the “United States in Congress Assembled.”

Regardless of the name by which this congress was known or its exact power or structure, each of the colonies/states had one vote (regardless of the number of delegates it sent). When the government under the Articles of Confederation proved ineffective and the Constitutional Convention was debating the composition of the legislature to be proposed, larger states urged that voting power be proportional to population (the Virginia Plan), while smaller states sought to retain equal voting power (the New Jersey Plan). Ultimately the convention agreed to the Connecticut Compromise, which provided for two houses in the new Congress: the House of Representatives (states represented in proportion to population) and the Senate (each state choosing two senators).

This two-house structure for a legislature is called bicameralism, from the Latin word camera, meaning “chamber” or “room.” (In this context the words “house” and “chamber” are synonymous.) The Congress under the Constitution was of course not the first bicameral legislative body. Not only did the British Parliament (whose authority over America was denied by the Declaration of Independence) have two chambers, the House of Lords and the House of Commons, but two houses were the usual, if not the sole, manner of organization of colonial, and later state, legislatures.

As the Connecticut Compromise shows, bicameralism allows for different methods of selection or representation. (Until the Seventeenth Amendment, senators were chosen by state legislatures rather than popular vote.) Until the reapportionment cases of the 1960s, some states used factors other than population equality to apportion seats on one house of their state legislature. In New Jersey,for example, each county was formerly represented by one senator, until the rejection of the so-called “federal analogy” by Reynolds v. Sims, 377 U.S. 533, 84 S. Ct. 1362, 12 L. Ed. 2d 506 (1964), eventually resulted in the current structure, in place since the 1973 election: 40 legislative districts, with substantially equal population, each electing one senator and two members of the general assembly.

It is this method that is described by the title of this post. Each district is represented by legislators elected by the same voters (and, in three out of five elections, at the same time). The result (encouraged by the gerrymandering used after each census to create the districts) is that almost every (if not every) district is represented by three people from the same party. What, then, is the point of having two houses? In a future post I will suggest an alternate method of selecting senators that might answer this question.