In a column posted on this blog earlier this week, Michael Dorf pushed back against what he called “the current bout of Senate skepticism”—in particular, against complaints about the system of electing two senators from each state. Part of Dorf’s argument is aimed at the idea that the structure of the Senate systematically favors Republicans: he argues that the Republican slant is neither as steep nor as permanent as some of the critics believe. I tend to agree with Dorf on those questions, albeit not completely, and in any event I’m going to leave that part of the conversation aside. My interest here is in a more fundamental part of Dorf’s argument, one that goes to the merits of the Senate without respect to partisanship. It’s about the connection between the Senate and federalism.
Dorf argues that in a federal system with a bicameral legislature there are plausible reasons for allocating representation in one house on the basis of the polity’s component political units. So if it is desirable for the United States to have a federal system, he writes, then the Senate might be a necessary part of a package that is worthwhile overall. In his words, “American federalism has benefits as well as costs, and insofar as the Senate is an accommodation for federalism, it accommodates the benefits as well as the costs.”
Dorf is among the sharper analysts of constitutional law, and I usually find his ideas persuasive. But my purpose in this post is to push back against the idea that the Senate is justified as part of the package deal that gives us our federal system. To be clear, Dorf might not actually be saying otherwise: his analysis is carefully conditional. But I think readers will reasonably take away the idea that federalism makes sense of the Senate, and in any event many Americans hold some form of that view. So it seems worth taking this opportunity to point out that there is a serious problem with thinking that federalism justifies the Senate.
The central problem, I suggest, is simply that it is a mistake to think that the system for apportioning senators makes any significant contribution to American federalism. Nothing valuable about our federalism depends on letting each state have the same number of representatives in the Senate.
American federalism isn’t perfect, but it’s enormously valuable. Insofar as federalism is about letting power be exercised locally rather than centrally—which is half of what federalism is about[1]—federalism has a familiar basket of virtues. It permits local variations in policy, which can both increase the degree to which law reflects the preferences of the people governed and furnish possibilities for experimentation and comparison. Vesting significant government responsibility at the state and local levels sometimes fosters more responsive government than national government would, and it also sometimes increases government’s ability to tailor policies to local conditions. Moreover, locating a large share of policymaking at the state level healthily reduces the stakes of what might otherwise be winner-take-all national politics. These are other virtues of the localist aspect of federalism are real, and important, and worth preserving.
But the fact that there are two Senators from each state does basically nothing to preserve or facilitate federalism. It might feel like a federalism sort of thing, because it’s about states and because it has historical roots in a negotiation among states. But the feeling is misleading. The substance of federalism is the division and sharing of power among the national and state governments, and nothing about those allocations of power depends on having a Senate with the same number of representatives from each state.
To see why, imagine a bicameral Congress in which the Senate was exactly like the Senate we know, except that senators were apportioned by population. There would be 100 senators, elected for staggered six-year-terms. (There’s value in having two houses, and there’s value in having the two houses be of different sizes, and there’s value in having a house with longer terms and a house with shorter terms. None of those things depends on each state’s electing the same number of senators.) The sole change to the Senate as we know it would be that instead of each state’s getting two senators, each state below a certain population would elect one senator only, with the balance of senators distributed among the larger states in proportion to population. Or maybe each individual senator would represent a district of a little more than three million people. (In which case not every state would have its own senator. Some states would inhabit multistate senatorial districts, much as states today inhabit multistate federal circuits for judicial purposes, while remaining separate in every other sense.) Suppose this change in apportionment were put into practice. What exactly would change about the allocation of power between the national government and the state governments?
Probably nothing. The constitutional powers of the federal government would have not been altered in any way, nor would those of the state governments. State governments would continue to be elected by their local populations and to have budgetary and policymaking power not derived from national authority, just as is the case today. State governments would continue to be the primary policymakers in education, crime control, land use, and all the other areas in which state governments are now the primary policymakers. The entire bundle of federalism-based judicial doctrines, from anticommandeering to sovereign immunity to dormant commerce to the various forms of statutory preemption, would, for better and for worse, be exactly the same. In short, federalism—the division and sharing of power between the national and state governments—would remain undisturbed.
To be sure, there would be resulting changes in the substance of federal law, and those changes would be more in keeping with the preferences of some states’ voters than those of other states. Fewer federal dollars would go to North Dakota, and more would go to Texas. The national government would probably spend more on mass transit and less on agricultural subsidies. The policy priorities of the national government would change. But the architecture of federalism would not. Congress would not gain the power to instruct state legislatures to legislate, and federal courts would continue deferring to state-court interpretations of state law. California would have an income tax, but Florida wouldn’t, and Chapter 2 of the UCC would still not be the law in Louisiana. More generally, nothing about the changed composition of the Senate would reduce the ability of state governments to make their own diverse and locally chosen policies in all the areas where state governments make policy.
Once upon a time, it might have made sense to say that changing the basis for electing senators would change the dynamics of federalism itself because senators acted as the particular representatives of state governments. When senators were chosen by state legislatures, they might have been more attentive than their colleagues in the House of Representatives to the interests of state governments, as differentiated from the interests of the population generally. A decisionmaking institution within the federal government that was distinctively attuned to the needs or perspectives of state governments as such might indeed have done work in shaping or maintaining the architecture of federalism. But any distinct function of this kind that the Senate executed once upon a time disappeared long ago with the move to popular election of senators, first as a practical matter by the middle of the nineteenth century and then as a formal matter with the Seventeenth Amendment. In modern America, there is no evidence that the voting behavior of senators is affected by their ostensibly being the representatives of states as such. Grade schoolers are still taught that members of the House of Representatives represent “the people” and senators represent “the states.” But in practice, a senator of Party P and State S votes pretty much like a Representative of Party P and State S. Controlling for partisanship and location, that is, a senator has no greater tendency than a representative to cast votes sensibly coded as “federalism-protecting” or “protecting the prerogatives of states.”
All in all, it is hard to see why twenty-first century federalism requires, or is in any meaningful way facilitated by, a Senate with two senators from each state. If the practice of electing two senators from each state is defensible (in any sense beyond “that’s the rule,”) it will have to be for other reasons.
So far, my analysis has proceeded on the understanding that the idea that there is a federalism-based justification for giving each state equal representation in the Senate means that that system of representation somehow assists in the functioning of federalism as a going concern. (Which it doesn’t, as explained above.) But there is also another way of thinking that the Senate’s composition should be understood as part of the overall package of American federalism, which we must take or leave in gross. It’s historical, rather than functional: without the agreement creating a Senate with two members from each state, the Constitution would not have come into existence at all. The system’s origin lies in the coming together of states, and the method of apportionment is a consequence of that origin. So if we’re happy that the states function together as a union under the U.S. Constitution, then maybe we need to accept the system for electing senators as part of the cost of doing business.
I think that’s wrong. Not the historical claim: it’s perfectly reasonable to believe that without the compromise that produced the Senate as we know it, the Constitution would not have been adopted at all.[2] But that historical fact is not a reason to regard the present system for electing senators as reasonable on the merits, or even as something that we should accept as necessary, rather than thinking of it as the result of an unfortunate historical contingency that should be cured at the first possible opportunity. The Framers had to reach compromises about slavery, too. Slavery and the Senate stand at very different points on the continuum of injustice, but the point should be clear enough: a compromise can be required for getting to yes and also stink.
My sense that the compromise stinks is not based on any general principle that all legislative apportionments must strictly conform to the one-person, one vote model. (Indeed, there isn’t even any such thing as “that model,” because there are disputes about whether the basis of apportionment that one person one vote requires is residents, or citizens, or voters, or maybe something else.) Different federal systems have different aims and face different circumstances. So the best solution for one federal system might not be the best solution for every federal system.
Instead, my sense that the compromise stinks rests on two propositions about the specific system that is the United States. The first is that in this system, one person one vote is an important value, and rightly so. The second is that the present system for allocating senators, which deviates drastically from one person one vote, is not justified by anything other than the force of an unprincipled deal made under duress in 1787. Establishing that second proposition would require wrestling with more aspects of the problem than I have discussed here, and a blog post is no place to attempt a comprehensive argument. For present purposes, suffice it to say that it’s my view—and I know not everyone agrees—that none of the other functional or normative arguments usually adduced to justify the system of giving each state two senators makes much sense.
Equal representation for each state in the Senate was unprincipled when the small states extorted it in 1787—Madison was apoplectic—and it has been unreasonable ever since. From the day the Philadelphia Convention rose, we Americans have done our best to make up reasons why the arrangement is sensible, or even wise. Partly to resolve the cognitive dissonance of living under a system with a structural injustice built into it, we have tried to prettify, with stories about federalism and sovereignty and “representing states,” what was in reality no more than self-interested parties to a negotiation cashing in on their threat advantage. The prettifying stories do not withstand critical scrutiny. For rule-of-law reasons, the unprincipled deal extorted in 1787 remains in force, and I do not believe that either Congress or the judiciary has the authority to change it. But that doesn’t mean it’s justified in any other sense.
[1] The other half is about letting power be exercised centrally. But as applicable here, it’s the localist half of federalism that matters.
[2] I’m prepared to treat that proposition as true, though nobody knows for certain what would have happened in a counterfactual circumstance.