Next month Californians will decide whether they want to remove Governor Gavin Newsom from office before the end of his first term. This is the second time this century that Californians have considered recalling their governor; in 2003 voters removed from office Gray Davis (who was less than a year into his second term), and the same day installed Arnold Schwarzenegger as the state’s chief executive. And as was true that time around, some folks today are arguing that California’s recall voting rules violate the equal protection rights of the incumbent governor and his supporters. But just as in 2003, this constitutional challenge is wide of the mark. California’s voting process might be unwise or needlessly confusing, but it is not unconstitutional in the way critics have recently charged.
California’s recall mechanism operates as follows: Voters are asked to weigh in on two separate decisions through votes tallied on a single ballot. The first decision for voters is whether to recall the incumbent. If a majority (or equal number) of voters reject the recall, the incumbent remains in office and no further voter preferences are consulted. But if a majority of voters end up favoring recall, then the results that voters register on the second (essentially contingent) part of the ballot are consulted to determine who will fill the recall-created opening.
How, according to critics, does this violate equal protection? In an op-ed in the New York Times last week, prominent constitutional scholar/commentator and Berkeley Law Dean (and personal friend of both our ours) Erwin Chemerinsky and his Berkeley Law colleague Aaron Edlin (a noted law and economics scholar) contend that California’s recall procedures run afoul of the one-person, one-vote equal protection mandate by weighing the preferences of some Californians more than of others. They write that because an incumbent governor who loses on the first part of the ballot (the recall decision) cannot under state law have his name appear on the second part (the vacancy-filling decision), his voters are treated unfairly in violation of the Constitution:
Imagine that 10 million people vote in the [upcoming] recall election and 5,000,001 vote to remove Mr. Newsom, while 4,999,999 vote to keep him in office. He will then be removed and the new governor will be whichever candidate gets the most votes on the second question. In a recent poll, the talk show host Larry Elder was leading with 18 percent among the nearly 50 candidates on the ballot. With 10 million people voting, Mr. Elder would receive the votes of 1.8 million people. Mr. Newsom would have the support of almost three times as many voters, but Mr. Elder would become the governor. . . [If Newsom loses on the] first question on the recall, [that] effectively disenfranchise[es] his supporters on the second question.
Chemerinsky and Edlin argue that such a scheme “violates a core constitutional principle that has been followed for over 60 years: Every voter should have an equal ability to influence the outcome of the election.” Pointing to the seminal malapportionment cases from the 1960s—in which the Supreme Court first announced the famous one-person, one-vote principle and required districts for legislative office (both federal and state) to be equally populous — Chemerinsky and Edlin contend that California’s approach to gubernatorial recalls flouts basic constitutional equality norms: “If Mr. Newsom is favored by a plurality of the voters, but someone else is elected, then his voters are denied equal protection. Their votes have less influence in determining the outcome of the election.” Indeed, Chemerinsky and Edlin contend that their assertion that California’s process is unconstitutional “should not be a close constitutional question.”
They also assert that “[t]his issue was not raised in 2003 before the last recall, when Gray Davis was removed from office after receiving support from 44.6 percent of the voters. But his successor, Arnold Schwarzenegger, was elected to replace him with 48.5 percent of the vote. So Mr. Schwarzenegger was properly elected. This time, we hope that a state or federal lawsuit will be brought challenging the recall election.”
Chemerinsky and Edlin are wrong on both counts: we see no voting inequality concern of constitutional magnitude here, and the argument they advance was indeed raised (by the incumbent Governor), refuted, and rejected during the Gray Davis recall contest. Indeed, one of us (Amar) made both of these points more than 15 years ago, first in online academic commentary in the months leading up to the Davis recall event, and then again (ironically enough) in the pages of the California Law Review not long thereafter.
On the constitutional merits, the fundamental and fatal flaw in the Chemerinsky/Edlin thesis is their unarticulated and undefended characterization of the two parts of the recall ballot as constituting a single election in which voters are being asked whom, out of the entire field of wannabe governors including Gavin Newsom, they prefer. Yet the recall vote is no such thing. It is two separate voter decisions—that simply (and for efficiency’s sake) are being conducted by means of a single ballot.
Suppose California law were to disaggregate its two-step process. Imagine that state law provided that at Time 1 (t=1), there would be a recall election by which the incumbent governor would be removed if he were to receive less than 50% support, and that if this were to happen then sometime later at Time 2 (t=2), there would be held a second election in which the highest vote-getter (even if she earns less than 50% support) wins and becomes the new governor. And suppose further that California explicitly adopted a rule that any governor who is recalled at t=1 cannot run again in the next election at t=2, such that a recall operates as a short-term disqualification from gubernatorial office. The Chemerinsky/Edlin argument would suggest that if this process were in place today, the t=2 election would be unconstitutional because supporters of Newsom wouldn’t be able to vote for him again. Chemerinsky and Edlin are correct that Newsom supporters’ “votes would have less influence in determining the outcome of the [replacement] election [at t=2]”—indeed, their votes would have absolutely no influence at all.
But that is the expected and natural effect of Newsom’s being disqualified from consideration in the replacement election, an effect that Chemerinsky and Edlin don’t even mention, much less deal with. Ballot-access rules often mean voters who support a particular person won’t have their votes counted equally (indeed at all), and yet that doesn’t necessarily violate equal protection. Indeed, there are many ballot-access rules that disqualify would-be candidates from being considered and potentially elected by voters in particular elections at particular times. These include age, residency, and citizenship requirements, as well as bans on office-holding arising from impeachment and conviction by legislative bodies. (Indeed, in Amar’s last column he discussed precisely that topic, in the context of former Illinois Governor Rod Blagojevich’s frivolous recent litigative attempt to undo the lifetime ban on state officeholding imposed on him by the Illinois Senate after his impeachment process over a decade ago.)
Consider another perfectly permissible device that functionally denies voters the opportunity to elect whomever they might want: term limits. States can and do impose term limits on state executive and legislative offices, and no one seriously thinks that violates the federal Equal Protection Clause. Suppose California said that its governor can “be elected to the office only once” (somewhat similar to the federal Constitution’s Twenty-Second Amendment’s provision that “[n]o person shall be elected to the office of President more than twice”). Under this scenario, an elected governor who was recalled before the end of her term could not be considered by voters during the election to fill the office no matter how much voter support she might have.
There is a long list of other well-accepted rules that do not violate the Equal Protection Clause but that prevent someone from being considered for office no matter how popular she may be relative to other contenders. One example very similar to the recall question involves so-called “sore loser” laws that exist in almost all states, which prevent a losing candidate from a major party primary from running in the subsequent general election as the nominee for a minor party or as an independent candidate. Or consider sophisticated voting protocols (like that used in the 2018 congressional election in Maine) by which candidates are removed in early rounds of ranked-choice regimes and thereby excluded from further consideration, even though some of these early losers may in theory be more popular, head-to-head, than the ultimate victor.
All of these ballot-access restrictions, and others too, effectively mean someone who wants to be elected and might be “more” popular among the voters than the ultimate winner is nonetheless denied a path to the office. Yet so long as a state chooses non-discriminatory, non-partisan, and reasonable candidate eligibility criteria, these restrictions should be—and routinely are—upheld by courts. In the present case, a legitimate and reasonable justification for excluding Newsom (or Davis) from the second part of the ballot is easy to imagine. By disqualifying a recalled incumbent from the “replacement” election (through the second part of the ballot), thereby temporarily sidelining someone so contentious that he just got recalled from the State’s highest and most visible office, California might easily be seen as promoting harmony and stability for the State and ensuring the State survives intact whatever crisis brought on the recall. It may be true that should the incumbent lose on the first part of the ballot, he is treated differently on the second part of the ballot than are other candidates for the newly vacant governorship, but that is because he is relevantly different – he is the only person who just got recalled.
In our view, the interest in keeping a recalled person on the sideline for the next election (or beyond) is actually stronger than the interests in new blood that justify absolute term limits (which have been upheld as constitutional by the state and federal courts in California). And a voter-recall-imposed disqualification from officeholding may also be easier to defend on democratic terms than are bans on officeholding imposed after impeachment processes, insofar as the people themselves (rather than their elected representatives) directly bring about a recall. For these reasons, an equal protection claim along the lines Chemerinsky and Edlin suggest would be extremely unlikely to succeed, and rightly so—it simply fails to distinguish between (1) an election in which some qualified candidates (and their supporters) are unjustifiably treated worse than others, and (2) an election in which some wannabe candidates are validly excluded from consideration in the first place.
To be sure, this distinction may be harder to see in California’s recall process than in some of our other examples of disqualification because voters register their views as to the two different election steps of recall and replacement through a single ballot. But that makes no legal difference. Indeed, if one were to (wrongly) view California’s regime as a single electoral decision (“who should be the governor going forward”) rather than two separate decisions that California happens to collapse into a single election day and ballot, then all candidates could voice the equal protection complaint Chemerinsky and Edlin advance on behalf of Newsom supporters alone. Newsom (and his supporters) could claim, as Chemerinsky and Edlin do, that Newsom is disadvantaged in the second part because he is not on the ballot on that part. But supporters of Larry Elder and other wannabe successors could equally complain that they are disadvantaged in the first part because their preferred candidates are not allowed to “run against” Newsom even though these candidates might arguably be preferred to him. In other words, although Chemerinsky and Edlin may be right that Newsom alone has to get 50% of the vote to be the future governor whereas other aspirants (like Elder) could win with, say, 20%, on the second part of the ballot, it is also true that Newsom could thwart the recall and hence be the future governor with 51% of the vote on the first part even if 80% of the electorate would prefer Elder in a head-to-head competition between the two. That is, just as Newsom is excluded from the second part of the ballot, all others (including Elder) are excluded from the first part. The two parts of the ballot effectively involve two distinct questions, so they are governed by different rules. And the distinct questions may well affect the voters’ decisions: even voters who in the abstract would prefer Elder (or another candidate) to Newsom as future governor might nonetheless vote against Newsom’s recall because they believe that any recall (even of a disfavored candidate) would be politically divisive, incite protests, or incur other costs that outweigh their perceived benefits from installing their preferred candidate for a partial term. Or they may vote against recalling Newsom simply because they dislike a subset of the other contenders even more, and don’t want to run the risk that one of that subset might replace him.
And if Elder supporters were to challenge his exclusion from the first part of the ballot, the natural and compelling response would be “Elder can’t complain about being excluded from the first decision because it’s not really an open all-comers election; it’s just a referendum on the incumbent.” But that reaction simply drives home that: (1) there are two separate electoral decisions on which Californians are asked to weigh in; and (2) some elections (as it happens, both Parts One and Two of the recall ballot) are not open to all comers. Part Two may of course be more wide open than Part One, in that many candidates meeting a relatively relaxed set of established qualifications can run. Yet the key question is whether recently being recalled can be a valid basis for exclusion from a subsequent (albeit very soon thereafter) replacement election.
As noted above, we think this basis for ballot exclusion—which is neither partisan nor irrational—would be easy for courts to uphold. Perhaps there is room for argument here, but it is an argument Chemerinsky and Edlin do not advance because their analysis doesn’t identify much less engage the crucial constitutional issue, namely that California’s recall process involves a ballot disqualification rather than (simply) a straightforward equal protection question.
All of this helps explain why the invocation by Chemerinsky and Eldin of the malapportionment cases is beside the point. In those cases, the candidates for whom voters expressed preferences were all eligible to be considered –no ballot-access issue was implicated there. When different voters within a state are voting for eligible candidates for the same office (e.g., state legislator or U.S. Representative), votes must be weighted equally. But people can’t bring a claim for violation of the one-person, one-vote principle when their votes for a candidate not eligible to serve (say, a candidate who doesn’t live in the state in which the election is held or a candidate who has been termed-out or impeached) are not counted equally, or indeed not counted at all.
Not only does the Chemerinsky/Edlin argument miss the key question of ballot access, but it wrongly asserts that the theory they advance is novel and has not been raised in litigation. In fact, as Amar analyzed in some detail in the writings mentioned above, Governor Gray Davis on August 4, 2003 filed a lawsuit in which he invoked the Fourteenth Amendment to ask the California Supreme Court to, among other things, rule that if a majority of voters were to opt to recall him a few months hence, the federal Constitution still entitles him to be a candidate on the second part of the ballot – the part that asks voters to fill the recall-created opening. That is, he made precisely the same equal protection argument Chemerinsky and Edlin assert. The California Supreme Court (rightly) denied relief then. And we are confident that it (or the U.S. Supreme Court) would do the same today.
As an aside, Chemerinsky and Edlin suggest that Gray Davis’s claim was weaker than the one Gavin Newsom has today because Schwarzenegger got more votes than Davis in 2003 whereas Newsom looks more popular relative to the 2021 field. That suggestion is also flawed, for a number of reasons. First, just because Schwarzenegger got more votes on Part Two than Davis did on Part One doesn’t mean that Davis couldn’t have gotten more votes than Schwarzenegger on Part Two had he been eligible to appear on that Part. Consider, for example, that some Davis supporters might have thought that the recall vote was sure to go against Davis on Part One and for that reason didn’t go vote at all. Had they known their support for Davis might have helped him on Part Two, they very well might have shown up and vaulted him above Schwarzenegger on Part Two even if Davis lost on Part One. Or perhaps some voters who did show up to vote may have wanted to recall Davis to send a message but still would have preferred then to put him back in office (duly chastised) to electing Schwarzenegger. Moreover, if a court had agreed with Davis on the merits after the election but felt it could not impose an aggressive remedy of ordering him reinstalled in office because he received fewer votes than Schwarzenegger, the court could still perhaps have remedied the equality violation by providing some kind of declaration (helpful in the future) and nominal damages. So the rejection of Davis’s claim in 2003 made sense not (as Chemerinsky and Edlin might suggest) because a court couldn’t have properly provided any meaningful relief, either before or perhaps even after the election, but because Davis’s equal protection claim was wrong as a matter of constitutional law.
None of this is to say that California’s recall system is necessarily wise. It does risk the election of fringe candidates; for that and other reasons (as Amar has argued in earlier writings) there may very well be better approaches to the problem of unpopular governors. And even sticking with the current system, there may be less confusing ways to present the distinct electoral questions being asked. But the question here (as with term limits and other limitations on officeholding or ballot access) is not whether California’s approach is the wisest or clearest according to any analysts’ views, but merely whether it is constitutionally permissible.