Vanderbilt Law Review


Paul H. Edelman

First Page



"One person, one vote" sounds like a simple mathematical equation. Actually, it isn't quite that easy, but over the last forty years, the Supreme Court has distilled a fairly stable and predictable test for resolving the basic issue of equal representation: how much population difference between districts is permissible? In one area of representation, however, the Court has gotten the math wrong. In its only opinion on the decennial apportionment of Congress, the 1992 case U.S. Department of Commerce v. Montana, the Court punted. Rather than apply its well-established test from the districting cases, the Court deferred to Congress on the ground that different ways of measuring equality of representation produced different apportionments, and thus Congress, rather than the Court, should choose the best measure. Unfortunately, that conclusion was based on a mathematical error.

As an abstract matter, applying the districting test to apportionment makes sense. Apportionment and districting are opposite sides of the same coin. In districting, one has a fixed number of representatives, and the geographic area must be cut into pieces to accommodate them. In apportionment, the geographic area is already divided into states, and the representatives must be parceled out among the divisions. In both instances, the goal is "equal representation for equal numbers of people." In both instances, the same test should apply.

But when the Court tried to apply the districting test to apportionment, it was misled by its mathematical mistake. Relying on numbers provided by the parties, the Court thought it was looking at a calculation of relative deviation-the test used in the districting cases-when it was not. Instead, it was looking at a different computation, and, not surprisingly, this alternative computation produced results that conflicted with other indications the Court had. The Court thus concluded that the relative deviation test could not be applied to apportionment because the results that the Court reached under the test sent mixed signals about the appropriate measure of representation.

This Article does the mathematics correctly. It provides a unified account of "one person, one vote" for both the districting cases and the apportionment of Congress, explaining why the same measure of "one person, one vote"-relative deviation-should apply to both districting and apportionment. I argue that the Court may be constitutionally required to apply this test in the apportionment context, and I demonstrate the method of finding the apportionment that best satisfies the requirement.

The apportionment method that best satisfies the "one person, one vote" test is unlike any method previously in use. Calculated under this method, the current Congress would have three fewer representatives from California and one more from each of South Dakota, Delaware, and Montana. In general, larger states lose representatives and smaller states gain.

Part II describes the evolution of the "one person, one vote" standard for districting cases. Part III relates the history of apportionment of the House and examines the sole Supreme Court case on the requirements of apportionment. In particular, this Part shows how an arithmetic error caused the Court's confusion about how to measure compliance with "one person, one vote" in the apportionment context. Having established a clear foundation, Part IV sets forth the prima facie case that the current apportionment of the House is out of compliance with the "one person, one vote" test from the districting cases. Part IV also considers possible counterarguments and concludes that they are not persuasive.

This does not end the inquiry, however, since under the "one person, one vote" test, the government has the opportunity to demonstrate that the current apportionment serves a legitimate political interest. Such a defense relies on technical properties of apportionments, and so I begin Part V with a discussion of the various methods of apportionment, including the new method that I propose. After laying the groundwork, I show that the legitimate government interests that support apportionment do not support the use of the current method of apportionment over my proposed method. In Part VI, I outline some of the political consequences that would result from my proposed method of apportionment. .