Fixing the Senate
A major contributor to our nation’s political dysfunction involves the disproportionate control by small states over our congressional lawmaking processes. A minuscule percentage of our electorate holds dramatically outsized influence. Although seats in the House of Representatives are, for the most part, allocated on a population basis, seats in the Senate are not. Each state has two Senators without regard to relative differences in population.
It is well known that the Senate is an historical anachronism, dating to the framing period. The Constitution’s original architects negotiated a tradeoff. States agreed to relinquish important aspects of sovereignty in exchange for avoiding the manifold defects operating under the Articles of Confederation. This was part of the larger project of forming a “more perfect,” even if still deeply-flawed, union. That compromise, which resulted in affording seemingly disproportionate power to small states as compared with a population baseline, seemed to many reasonable at the time. This was especially so given the even greater limits that states as states held over federal law and policy making under the Articles and given the perceived risks that the larger states might run roughshod over smaller states in formulating federal policy under any new constitutional plan. Notably, this related to the highly sensitive and contested matter as slavery, an issue that ultimately shook our nation to its very core, resulting in the Civil War and the Reconstruction Amendments, the aftermath of which continues to hold major contemporary repercussions. In the original design, state legislatures were authorized to select the state’s senatorial delegation, but the Seventeenth Amendment, ratified in 1913, changed that to have states choose their Senators by popular election.
Even with popular elections, the constitutional compromise of two houses, one based on population and the other based on statehood, generates a host of modern problems. The most notable is a virtual stranglehold on lawmaking. The filibuster, which requires 60 of 100 votes to enact most laws, exacerbates the problem. As few as twenty of the least populous states hold the power to block legislation in the Senate. A simple calculation from these state population data demonstrate that 14.72 percent of the population reside in the lowest-populated 25 states, and 10.19 percent of the population reside in the lowest-populated 20 states. Setting aside the filibuster, some might consider it morally problematic that 14.72 percent of the population can effectively block desired federal policy. And yet, this dramatically understates the difficulty. The filibuster appears to lower the relevant population percentage to 10.19, and following a simple insight derived from the famous book by James Buchanan and Gordon Tullock, The Calculus of Consent, the actual figure cuts even that roughly in half. A simple majority of voters within a controlling block of states—now twenty—hold the theoretical power to block federal policy. This means that the real stranglehold belongs to a mere 5.2 percent of the population. It is difficult to imagine any credible moral justification that allows such tight control over federal policy by such a tiny percentage of the electorate.
Although the problem is fairly easy to identify, it might seem harder to fix. Having just finished grading a set of exams, I have gained an insight that until now had eluded me. After some pencil and paper calculations, I have a fairly simple, and I think fair, proposal. First, allow me to explain the methodology.
My goal is to devise a solution that is minimally disruptive of the overall constitutional scheme. My preference for a modest tweak over a complete overhaul follows from both practical and theoretical intuitions. It is easier as a practical matter to change one aspect than to entirely reform our constitutional structure. And because the Constitution has so many interlocking features, it is likely to be far less disruptive to fit changes within the overall existing scheme. With those premises in mind, I set about devising a senate with the following features. First, it will continue to have close to 100 members. This will ensure that the body is appropriately deliberative and thus less susceptible to the necessary top-down procedural controls that are an inevitable feature of the more populous—435 voting members—House of Representatives. Second, no large state will have more than double the present number of senators, and no small state will have fewer than half the existing senators. This is a value judgment, and to that extent it is arbitrary. Indeed, it is arbitrary in the very same sense as is imposing final grades upon a distribution of exam scores. At some point toward the top and bottom, we group students together, and we do the same for cohorts in between. By convention, we tend to call these A+, A, A-, B+, etc., but where those lines are drawn, and how far up or down along the grading scale we go, is inevitably a value judgment, generally informed by past experience or institutional norms. Similarly here, past practices and the role of the Senate within our larger constitutional structure inform the judgments I have made.
Given these parameters, and assuming each state retained no fewer than a single senator regardless of population, the challenge in making the numbers work was defining the relevant population increments that would justify awarding additional senators above one. My exam-grading insight was that we could take the difference between the largest population state (California) and smallest population state (Wyoming) and derive a Total Population Value (TPV). Based on 2018 populations, the math works as follows: (California) 39,776,830 minus (Wyoming) 573,720 equals (TPV) 39,203,110.
The next step is a bit trickier. With this listing of states in descending order of population, I needed to identify what I will now call an Incremental Population Value (IPV). The IPV awards an additional senator to every state with a population that is greater than the IPV, while ensuring one senator to every state with a population below the IPV. The goal was to find the IPV that would create a senate as close to one-hundred members as possible without giving any state more than four senators (double the present number) or lower than one senator (half the present number). Working within these parameters, I took the TPV as the numerator, and divided it by 2, 3, 4, etc. as the denominator. Each calculation offered a possible IPV that could be used to calculate a senate, with the goal of getting as close as possible to a 100 members.
And so, I divided the TPV starting with 2 and continuing until 10, at which point I was able to successfully complete the task. The calculated IPVs, based on the increasing denominators, and correspondingly smaller fractions, are as follows: 2 (19,601,555), 3 (13,067,703); 4 (9,800,776); 5 (7,840,622), 6 (6,533,852); 7 (5,600,445); 8 (4,900,339), 9 (4,355,901), 10 (3,920,311). Again, under the formula, each state begins with a single senator, and for each IPV, it receives an additional senator, subject to a maximum of 4 senators total. After playing with these different IPVs, I realized that the IPV of 3,920,311, resulting from dividing the TPV of 39,203,110 by 10, provides the mathematical solution.
This results is the following 96-member Senate:
States with Four Senators (6 x 4 totals 24): CA, TX, FL, NY, PA, IL
States with Three Senators (6 x 3 totals 18): OH, GA, NC, MI, NJ, VA
States with Two Senators (16 x 2 totals 32): WA, AZ, MA, TN, IN, MO, MD, SI, CO, MN, SC, AL, LA, KY, OR, Ok
States with One Senator (22 x 1 totals 22): CT, IA, UT, NE, AR, MS, KS, NM, NE, WV, ID, HI, NH, ME, MT, RI, DE, SD, ND, AK
This senate produces some interesting numerical correlations. The six four-senator states comprise 41.24 percent of the population, and with 24 senators have 25 percent senatorial representation. California, with 12.13 percent of the nation’s population has the same representation as Illinois, with 3.91 percent. This is akin to grouping students together at the high end with a single grade, A or A+, yet failing to distinguish the student at the absolute top. Illinois benefits from an arbitrary line, and California suffers from the decision to limit the maximum senatorial delegation to four. The six three-senator states comprise 18.35 percent of the population, and with 18 senators hold a near-precise 18.54 percent of the senate. The sixteen two-senator states comprise 27.93 percent of the population, and with 32 senators, also a near-precise 28.26 percent of the senate. And finally, the twenty-two single-senator states comprise 12.24 percent of the population, and with 22 senators hold a still disproportionate 22.9 percent of the Senate. The not-quite-double representation as compared with population is akin to pulling up the bottom of the curve to hit a mean, in this instance, ensuring that no state has fewer than a single Senator without regard to population scarcity. In effect, the low population states benefit from grade inflation, with Wyoming, the lowest, holding .17 percent of the nation’s population, along with Connecticut, the highest, holding 1.09 percent, and with each holding 1.04 percent senate representation. (Because of the District of Columbia, percentages do not add precisely to 100).
In this scheme, more populous states continue to be underrepresented, whereas the low populous states remain overrepresented, but far less so at both extremes than under the present formula of two senators per state. This has the potential to blunt the power of very low population states to block federal policy against the will of overwhelming percentages of the electorate, while still affording disproportionate representation to smaller states.
To be sure, this is an imperfect compromise. I doubt anyone would devise this system if starting from scratch. But we are not starting from scratch. In addition, we are now seven states away from a constitutional convention. An obvious argument against my proposal is that lesser populous states would not go along. That might be true, unless, that is, the threat of a constitutional convention—something that has never in our history occurred—is perceived as a greater because it is our genuine constitutional unknown and because it is being considered at the peak of our national political dysfunction. If this proposal allowed for the possibility of improving our constitutional functionality without the need for, or risk of, a complete overhaul, that alone might justify supporting it as an incremental fix that could, depending on how it performs, become permanent. The proposal, interim or permanent, would require a constitutional amendment. If interim, it could include a sunset provision, allowing it to be tested, and then, if successful, it could become permanent with a subsequent amendment. Or it could simply be implemented at once.
I must note that Article V contains an arguably unamendable "equal suffrage" clause respecting the Senate. I am assuming that even that is subject to amendment, although this might require a special process to secure the acquiescence of small states. On this issue, scholars have taken opposing views.
Invariably after grading exams some students are less than satisfied with the outcomes. And sometimes students see things that the professor who wrote the exam missed. Professors are sometimes right, but not always. We all make mistakes. Here too, I might be overlooking something that eluded me but is obvious to others.
For that reason, as always, I welcome your comments.
[Special thanks to Michael Abramowicz helpful comments on an earlier draft.]