March 11, 2009
Written Testimony of Vikram David Amar
Associate Dean for Academic Affairs and Professor of Law, University of California, Davis School of Law
Before a Joint Hearing of the Committee on the Judiciary, Subcommittee on the Constitution, United States Senate, and the Committee on the Judiciary, Subcommittee on the Constitution, Civil Rights and Civil Liberties, U.S. House of Representatives
March 11, 2009
Chairman Feingold, Ranking Member Coburn, Chairman Nadler, Ranking Member Sensenbrenner, Members of the two Committees:
My name is Vik Amar, and it is my distinct honor and pleasure to be here today to talk with you about S.J. Res. 7 and H.J. Res. 21 - a proposed constitutional amendment introduced by Chairman Feingold concerning Senate vacancies. For over 20 years, beginning with my days as a student at the Yale Law School, I have been studying and writing about the U.S. Senate and the central roles it plays in our constitutional scheme. My Yale Law Journal student Note, "The Senate and the Constitution," looked at the ways the Senate was more central in the area of constitutional interpretation than even the Supreme Court. One of my tenure pieces at the University of California undertook a structural examination of the Seventeenth Amendment to uncover some unobserved consequences of direct election of Senators, and in one of my most recent law review articles I analyzed the reasons the Seventeenth Amendment prefers Governors to state legislatures and other bodies when it comes to temporarily filling Senate vacancies, which led me to question the constitutionality of Wyoming's vacancy-filling statute - a statute that delegates to political party leaders the task of compiling lists from which a replacement Senator must come.
I. Specific Reasons to be Cautious About Amending the Constitution in this Area and What a Prudent Constitutional Amendment Would Make Sure to Include
So my interest in and thinking about Senator Feingold's proposal and things like it go back a long ways. Let me begin by making clear I fully agree with the premise of the proposed constitutional amendment: there is ordinarily no better way to pick Senators than through popular election. While some modern scholars and analysts might question whether the Seventeenth Amendment (and the historical practice of increasingly widespread direct election in many states that preceded it) was, on balance, a good thing, I am not among them: any lamentable reduction in state governmental clout in the federal government occasioned by the move away from state legislative selection is more than offset by the populist virtues of direct election. So if the question were simply whether the people are better than both the state legislatures and state Governors at picking Senators, my answer would be an emphatic: "Yes."
A. The Problem of Extended Vacancies
But there are problems with eliminating temporary appointment power altogether. The first difficulty arises because elections take time. As the Continuity of Government Commission reported in 2003, "under ideal circumstances, states could hold elections within two months [of an unanticipated House or Senate vacancy] if they dispensed with party primaries and drastically accelerated other aspects of the campaign," but a more likely timeframe under real-world constraints might be three months. Three months doesn't sound like a long time, but such a delay in filling vacancies can matter a great deal when, as has been the case of late, the partisan balance in the Senate is close. This is especially true in light of modern filibuster practices and other supermajority rules and conventions. Very recent experience concerning the passage of this year's stimulus package highlights how even one vacant seat from Minnesota and/or one disability from the Massachusetts contingent can shape momentous legislation.
Delay in filling vacancies affects not only Senate actions, but also the states that are temporarily underrepresented and denied the equal suffrage in the Senate the Constitution takes extreme pains to guarantee. Indeed, the difficulty state legislatures experienced in promptly filling Senate vacancies was one of the key factors animating the move towards direct election that culminated in the Seventeenth Amendment. And the unusual representational structure of the Senate can magnify the unfair consequences of a Senate vacancy for a state. By many modern instincts, it is counter-intuitive enough that large states like New York and Texas should receive no more voice in the Senate than small states like Hawaii and Alaska, but the possibility that California should have half the voice of Delaware for any appreciable period of time in the Senate borders on the surreal.
The problem of vacancies lasting months is, of course, exacerbated substantially by the specter of terrorism in a post-911 world. As my University of Texas colleague Sandy Levinson has reminded, "[u]nfortunately, it is not fanciful to imagine an attack on Washington that would kill dozens of senators." The scary but not far-fetched prospect of, God forbid, a large number of Senators being killed or incapacitated such that a number of states or even parts of the country might lack Senate representation during the very months when key decisions about how the federal government must respond to crisis must be made should give every American pause before constitutionally eliminating all mechanisms for prompt if temporary replacement of fallen legislators. At a minimum, then, any constitutional amendment in this area should have a provision for a fallback mechanism that is triggered by some declaration of national emergency or some numerical threshold of Senate vacancy.
B. The Problem of Voter Turnout in Special Elections
A second problem of special elections, related to but beyond the question of delay, is the question of voter turnout. Although I have not undertaken an exhaustive empirical study, there seems to be a broadly shared and eminently plausible intuition that voter turnout when only one contest - even a U.S. House or Senate race - is on the ballot is likely to be much lower on average than when House and Senate races appear along with other state and/or federal office contests and/or ballot measures on a regularly scheduled election ballot. As NYU Professor Clayton Gillette has pointed out in an analogous context, "[i]t is . . . not surprising that voter turnout at special elections . . . is lower than voter turnout at general elections." To cite but one recent, if perhaps somewhat demographically unusual, example of seeming relative apathy in a special election, the voter turnout in the election held to fill only the U.S. Senate seat in Georgia last December was about one-half of the turnout in the regularly-scheduled November election just a month earlier -- and this low December turnout was despite the general understanding that the special election's results could determine whether Democrats would have a filibuster-proof majority in the Senate. The premise with which we began, that popular elections are the best way to pick U.S. Senators - a premise with which I agree - would seem to be most justified only when those popular elections are ones in which a broad cross section of statewide voters are encouraged and likely to participate. The turnout problem may also be more pronounced in some states than others, counseling caution when uniform federal mandates are being considered.
Of course, Senator Feingold's proposed constitutional amendment does not require special elections to fill Senate vacancies; it requires only that elections - special or regularly scheduled - be the exclusive means of filling such vacancies. But the longer a state waits to have a vacancy-filling election - either to save costs by consolidating the vacancy-filling election with an already-scheduled one and/or to increase voter turnout by combining the vacancy-filling election with other important decisions about which voters care - the longer the state (and the nation) must suffer the consequences of that state being under (or un- , in the case of a dual vacancy) represented in the "greatest deliberative body on earth."
C. Why State Practice and the Debates over the Seventeenth Amendment Demonstrate Governors Are Better than State Legislatures as a Fallback
Recognizing and balancing these concerns, almost all states have chosen to create temporary appointment power rather than use only elections to fill Senate vacancies. It bears noting that under the current Constitution, states are not obligated, but rather are merely authorized, to create temporary appointment power. And yet nearly all have. It is in the best tradition of federalism to recognize wisdom in the common practice of states.
If, then, as seems prudent, there should be some mechanism, either generally available or at the very least triggered by national emergency, for prompt vacancy-filling, we turn to the question of which branch of government is best suited to discharge the vacancy-filling power. Temporary gubernatorial appointment authority seems better than any of the alternatives. Governors are superior to state legislatures (and other bodies) here because Governors (unlike legislatures whose district lines are manipulated for partisan and other reasons) are elected by
and directly accountable to the exact same statewide electorate that elects Senators. Governors can also gather information privately about possible candidates and act quickly when time is of the essence. As Joseph Story said in this connection, "[c]onfidence might justly be reposed in the state executive, as representing at once the interests and wishes of the state, and enjoying all the proper means of knowledge and responsibility, to ensure a judicious judgment."
D. The Shape and Size of a Prudent Constitutional Amendment (Including a Provision Concerning House Vacancies)
To summarize thus far, I argue that any constitutional change that requires elections be held to fill vacancies contain, at a minimum, an exception that would authorize prompt gubernatorial appointments in times of national emergency. And indeed even outside emergency situations, if special elections were constitutionally mandated to be held within a specified time thought to be shortest practicable period necessary to organize them, it might nonetheless be advisable to retain gubernatorial appointment power to fill vacancies during the interim. After all, as noted above, even short vacancies can seriously prejudice underrepresented states as well as the nation as a whole.
Furthermore, if the Constitution were to be amended concerning Senate vacancies in these ways, I would recommend amending the provisions concerning House vacancies as well, to create a mechanism for prompt gubernatorial vacancy-filling power, at least in times of national emergency. Although vacancies of non-trivial duration in the House raise less severe democratic problems than do vacancies in the Senate, they are still undesirable. Constitutional amendments are invariably hard to pass and ratify; dealing with closely related problems in a single amendment is eminently reasonable, both in terms of constitutional structure and esthetics as well as enactment strategy. Providing for temporary gubernatorial appointment power in the House would undoubtedly require a constitutional amendment, since there is no provision in the current Constitution akin to the Seventeenth Amendment allowing anything other than elections to fill House vacancies.
II. How a Well-Crafted Statute Would Cure Most or All of the Perceived Defects in the Current System and Why Such a Statute Would be Constitutionally Permissible
That brings me to the question of whether improvements concerning the Senate (unlike the House) require any constitutional change at all. My own tentative view is that a statute along the lines of the bill promoted by Congressman Aaron Schock currently entitled the "Ethical and Legal Elections for Congressional Transitions Act (E.L.E.C.T.)" is the wisest course to pursue. That bill, in the tradition of the Continuity in Representation Act, would require that an election to fill a Senate vacancy generally be held within 90 days of the vacancy's creation, but would not disturb any existing state law mechanisms for a temporary gubernatorial appointment to be made during the 90-day period. The bill would also provide states some money to help defray the costs of special elections.
Although one might quibble with some of the proposed statute's details (including the choice of 90 days, rather than 120 days, etc.), I believe the basic approach is sound, and that statutes are superior to constitutional amendments in this area. A statute would be easier to enact than a constitutional amendment, and could also be more easily perfected in the coming years as more data is gathered based on actual experience in the states. In general, the only substantial reason to prefer constitutional amendment to statutory enactment would be to lock in the new legal regime and prevent Congress from subsequent legislative amendment or repeal. But I see no particular reason to distrust Congress in this particular area, and any subsequent statutory amendment would probably be an attempt to act on new information and would not likely represent an illicit Congressional effort to undo a worthy law.
An important issue becomes, then, would a law such as E.L.E.C.T. be constitutionally permissible? I think it would. Surely Congress can require vacancy-filling elections to take place within a certain period of days with respect to House vacancies (as in the Continuity in Representation Act) under its Article I, Section 4 power to "alter or make" regulations concerning the "time, place and manner" of federal legislative elections. There is no question but that a 90-day time-frame is a regulation of the "time" of an election the Constitution already requires states to hold. Nor does the fact that a time requirement would apply to vacancy-filling House elections rather than regularly-scheduled biannual House elections affect the analysis; Article I, Section 4, enacted at the same time as Article I, Section 2's requirement that Governors issue writs of election to fill House vacancies, textually speaks to Congress' power to regulate the time of all House and Senate elections.
The question of Congressional power over vacancy-filling Senate elections may seem a bit trickier. Certainly, Congress under the original Constitution had the power to regulate the timing of all Senate elections done by state legislatures, including elections done by state legislatures to fill unexpected vacancies. Indeed, Congress in 1866 passed an Act that regulated the manner and timing of all state legislative elections of U.S. Senators. The Act said that whenever there was a Senate vacancy of any kind, both houses of a state legislature, on the second Tuesday they were in session, must vote to fill the vacancy, and if no person was elected, both houses must continue to vote at least once each and every day thereafter of the legislative session.
Do the text and timing of the Seventeenth Amendment change any of this? I think the answer is "no." As for text, it is true that the last words of the vacancy-filling provision of the Seventeenth Amendment - "by election as the legislature may direct" - suggest that state legislatures enjoy discretion. To be sure, the phrase "as the legislature thereof may direct" or "as the Congress may direct" used elsewhere in the Constitution connote broad independence and leeway. For example Article II's use of the phrase "as the legislature thereof may direct" has been interpreted by the Supreme Court in Bush v. Gore as giving state legislatures extremely wide latitude in picking Presidential electors. But the key difference is that in the Presidential election context, state legislative discretion is not superseded by explicit Congressional power embodied in Article I, Section 4. Article I, Section 4 itself says state legislatures have power to prescribe times, places and manners - broad leeway - but that such power can be overridden by Congressional exercise. So even though the "as the legislature may direct" language of the Seventeenth Amendment connotes state legislative power, if that power is constrained by Article I, Section 4, then the Seventeenth Amendment provides no barrier to statutes like E.L.E.C.T.
But can we apply Article I, Section 4 Congressional power to a provision of the Constitution enacted after Article I was adopted? Grammatically we surely can. Article I, Section 4, speaks broadly of Congress' power to "alter or make" "at any time" the regulations concerning the time of "holding elections for Senators and Representatives" - not just some temporal or geographical subset of Senators or Representatives.
Moreover, everyone seems to agree that we can and do apply Article I, Section 4 to regularly scheduled (every six year) Senate elections held by the people of each state, even though these popular elections are created and provided for only in the Seventeenth Amendment, adopted after Article I, Section 4. And there is nothing in the text of the Seventeenth Amendment that distinguishes regular popular elections from vacancy-filling popular elections. If Article I, Section 4 applies to the former, it ought to apply to the latter as well, and there are no words in or legislative history of the Seventeenth Amendment to suggest otherwise.
Indeed, the legislative history strongly favors applying Article I, Section 4 to all of the Seventeenth Amendment's provisions. Southern Senators attempted, during the latter stage debates over the Seventeenth Amendment, to insert language that would have freed popular elections of Senators from Congressional control under Article I, Section
4. Although these attempts ultimately failed, the members of Congress who debated the matter at length seemed to assume and/or agree that without such language qualifying the Seventeenth Amendment, all of the popular elections it provided for would indeed be subject to Congressional Article I, Section 4 time and manner oversight. And even though the subjective understandings of the Amendment's drafters may not necessarily bind us today, their public proclamations of those understandings certainly informed what intelligent observers of the day likely understood the words to mean.
Finally, it bears noting that in the only other instance in which the post-1789 Constitution explicitly empowers states to do something they lacked power to do beforehand - the Twenty-First Amendment - the newly created state power is subject to preexisting federal legislative power to preempt. Section Two of the Twenty-First Amendment empowers states to create essentially federal laws concerning the in-state importation and distribution of alcohol, and yet the Supreme Court has held that this state empowerment does not abrogate Congress' Commerce Clause powers with regard to liquor: "The argument that "the Twenty-first Amendment has somehow operated to 'repeal' the Commerce Clause" for alcoholic beverages has been rejected."
For these reasons, E.L.E.C.T. is constitutional, and thus to my mind preferable, to changing the Constitution. And if there were any doubt about whether a statute such as E.L.E.C.T. might be struck down, a fallback severability clause could easily be added to the effect that if the requirement of a 90-day election were invalidated, then any state that chose not to comply with the 90-day timeline would lose not only federal funding for its special elections, but also federal funding for a large subset of its elections more generally. Although some care might be taken to comply with South Dakota v. Dole, it seems very likely a statute whose funding conditions would pass muster could be written in a way so as to encourage every state to comply with the 90-day time frame.
1. Vik D. Amar, The Senate and the Constitution, 97 Yale L.J. 1111 (1988).
2 Vikram David Amar, Indirect Effects of Direct Election: A Structural Examination of the Seventeenth Amendment, 49 Vand. L. Rev. 1347 (1996).
3 Vikram David Amar, Are Statutes Constraining Gubernatorial Power to Make Temporary Appointments to the United States Senate Constitutional Under the Seventeenth Amendment? 35 Hastings Const. L. Q. 727 (2008).
4 Continuity of Government Commission, Preserving Our Institutions: The First Report of the Continuity of Government Commission 19 (Washington, D.C., 2003).
5 See George H. Haynes, THE SENATE OF THE UNITED STATES: ITS HISTORY AND PRACTICE 79-117 (1960).
6 Sanford Levinson, Political Party and Senatorial Succession: A Response to Vikram Amar on How Best to Interpret the Seventeenth Amendment, 35 HASTINGS CONST. L. Q. 713, 723 (2008)
7 Clayton Gillette, "Direct Democracy and Debt," 13 Journal of Contemp. Legal Issues 365, 386 (2004).
8"Chambliss wins Second Term in U.S. Senate," ATLANTA JOURNAL CONSTITUTION, Metro Section, Dec. 2, 2008, viewable at:http://www.ajc.com/metro/content/metro/stories/2008/12/02/georgia_senate_runoff.html?cxntlid=homepage_tab_newstab.
9 See generally Amar, supra note 3.
11 Joseph Story, COMMENTARIES ON THE CONSTITUTION 264-65 (Hillard, Gray and Company, 1833). One intriguing possibility is that the Constitution could be amended such that a Senator herself would be required to name (and advertise to voters before she is elected) her would-be successor(s) (in rank order), who would then fill the seat should she depart the Senate before her term ends. Such apostolic succession (as we have in effect in the White House) is an intriguing possibility, but also one that might be unduly complex and/or create some opportunity for gamesmanship; such an approach would warrant much more study before adoption.
12 And if there were a concern that temporary appointees had too much incumbency advantage in the subsequent election, provision could be made to the effect that anyone appointed by a Governor to fill a Senate vacancy would be ineligible to be voted on for the Senate at the next election
13 For a discussion of this Act, see Haynes, supra note 4, at 85. See also, Robert Byrd, VOLUME I, THE SENATE, 1787-1989: ADDRESSES ON THE HISTORY OF THE UNITED STATES SENATE 392 (1988).
14 U.S. CONST., amend. XVII, § 2. In this regard, it might be worth noting that the original Constitution directly empowered state governors to make temporary appointments, whereas the Seventeenth Amendment authorizes state legislatures to empower state governors, an option known to but rejected by the original framers in 1787. See Story, supra note 10, at 264.
15 U.S. CONST., art. II, § 1, cl. 2.
16 U.S. CONST., amend.XXIII § 1.
17 531 U.S. 98 (2000).
18 See Haynes, supra note 4, at 108-115; Byrd, supra note 12, at 400-402.
19 Granholm v. Heald 544 U.S. 460, 486-487 (2005) (citing, among others, Capital Cities Cable, Inc. v. Crisp, 467 U.S. 691 (1984); California Retail Liquor Dealers Assn. v. Midcal Aluminum, Inc., 445 U.S. 97 (1980).
20 One potentially helpful thing a constitutional amendment could do that a federal statute could not is prevent a temporarily appointed Senator from running for the Senate seat in the next election. Article I, section 4's time and manner power would not extend to regulating the qualifications to be elected.
21 483 U.S. 203 (1987).