Count Every Vote -- All 538 of Them

Regular readers may remember my various posts last spring about my in-progress paper defending the Electoral College on pragmatic/procedural grounds. Well, it's finally done, and I've uploaded it to SSRN, where it's now  available for download.

Here's an excerpt from the introduction:

In the minds of many observers, the 2000 presidential election was the ultimate argument against the Electoral College.  For the first time since 1888, the popular-vote winner lost the electoral vote, and the people's second-favorite candidate became president.  Adding insult to injury, the electoral vote was itself extremely close, and the decisive state was even closer.  Thus, for the first time since 1876, an inversion between the electoral and popular votes coincided with a protracted dispute over the Electoral College winner—and for the first time ever, the dispute was settled not by Congress or a congressionally appointed body, but by the United States Supreme Court.  All things considered, the 2000 election became Exhibit A in many arguments for the elimination of the Electoral College.

It is possible, however, to take precisely the opposite view: that the 2000 presidential election was the ultimate argument for retaining the Electoral College.  This position is not sexy or politically correct; it is rooted in bald practicality, not democratic idealism.   Yet it has considerable merit.  A national popular vote system may be a good idea in theory, but in reality, it would become a nightmare in a close election.  And the interstate-compact system, a bizarre extraconstitutional hybrid that is neither a true popular-vote election nor a functional Electoral College, is even more nightmarish than direct popular-vote plans.  ... [I]t creates unclear lines of decision and a high likelihood of lengthy and unpredictable legal disputes when the winning candidate's margin is narrow.

These problems cannot be ignored, because without question, margins will sometimes be narrow.  In 2000, for example, Al Gore's popular-vote margin over George W. Bush  was barely one-half of a percent of the votes cast, which would be close enough to trigger a recount in some states.  Judge Posner acknowledged this point in a review of the Bush-Gore dispute, noting that "if Presidents were elected by popular vote, a nationwide recount might have been unavoidable in 2000 because Gore's margin was so small."   Yet that race looks like a landslide compared to Kennedy's 0.16 percent advantage over Nixon in 1960,  a hairbreadth edge that would have been very difficult to accept as final without the availability of a national recount to confirm the result. Even closer was the election of 1880, in which James Garfield defeated Winfield Hancock by just 9,070 votes, or 0.098 percent.   

In all three of these elections, and every other close vote in American history, the Electoral College isolated any controversies to individual states.  But under an interstate-compact or direct-election plan, such an eventuality would create controversy and legal wrangling on an almost unimaginable scale.  Every alleged anomaly, everywhere, would become the subject of potential legal action, because it would no longer matter whether a specific state was itself close.  And the incentive to look for anomalies would be extremely high: as Posner wrote in reference to the 2000 race, "There is little doubt that if Bush's people nosed around heavily Democratic precincts throughout the nation they would come up with colorable arguments about voter and tabulation error that might have determined the election [if the president were elected by popular vote]." 

It is worth recalling that Florida was hardly the only state where voting and vote-counting anomalies may have occurred in 2000. There were alleged anomalies in many other states,  but they were generally not litigated because those states were won in landslides, so the possible discrepancies did not matter to the candidates.  In a national popular-vote system, by contrast, the concept of "winning" a state would be meaningless; any close presidential election would be a nationwide free-for-all.  Every state and county courthouse in the country would become the front line in a pitched battle over a razor-close vote. 

Worse, the "final" resolution might not be so final, thanks to the likely thicket of legal challenges.  In an interstate-compact system, these could include a to the overall structure of the compact system itself. In a true direct-election system, finality could be an even bigger issue, as there would be no built-in mechanism to definitively end it all — no meetings of the electors at which the 538 votes that matter are finally cast, no congressional tally of those votes to conclusively choose the president and vice president. Without such a mechanism, challenges could drag on indefinitely, with state and federal judges extending deadlines in the name of "counting every vote."  What if "every vote" is not counted to everyone's satisfaction by January 20? What if differing interpretations of the vote count produce different "final" results, perhaps with different winners?  Which "winner" is inaugurated?

If the dispute over Florida's election in 2000 proved anything, it is that accurately measuring the will of millions of voters is very, very difficult.  Again quoting Judge Posner, "the counting of millions of ballots by any method is liable to error."  And yet Florida's vote total was just under 6 million;  the difficulties would become far greater if we needed an accurate national count of more than 100 million votes.   Many steps can and should be taken to improve the accuracy of our voting system, regardless of whether the Electoral College is retained.  But we must be realistic: there will always be the potential for disagreement about the result when the vote is close enough, and our electoral system must be robust enough to retain its legitimacy, functionality and timeliness even when such disputes occur.

As measured against the national popular vote, the Electoral College will very occasionally, in a very close election, give us the "wrong" president.  But it always gives us a president, and does so in a reasonably timely fashion, well before the January 20 inauguration. It may be unfashionable to defend a system that sometimes fails to accurately reflect the national popular will, but as long as inversions of the electoral and popular vote remain rare and limited to very close elections — ones in which the popular will is hardly overwhelming, and may indeed be quite unclear — this admittedly imperfect system is vastly preferable to the alternative.  Whatever its abstract merits, the practical reality is that direct election by national popular vote could, in a razor-close election, produce a true legitimacy crisis that would threaten to seriously undermine the presidency itself.

I've submitted the paper (or, ahem, "article") to a whole bunch of law reviews, law journals, etc., hoping to get it published. Whether I get any nibbles remains to be seen. If I fail, I'll make some revisions and try again in the spring. But I wanted to give it a shot now.

I'm told it's okay to post the article on SSRN while it's under consideration for publication elsewhere. Apparently law journals, in contrast to some other types of publications, don't generally care about that -- and indeed, apparently an SSRN posting can occasionally generate enough buzz to make it more likely that a paper will get published. I'm not expecting that, however. I mostly wanted to post it there so that y'all, my blog readers, can finally look at the darn thing, considering how many of you asked back in April and May when you'll be able to read it. :)

Many thanks to Professor Mayer, who has been incredibly helpful throughout this whole process. A footnote on Page 1 thanks him, and also thanks my dad, "Joe Loy, a retired elections administrator in the state of Connecticut, who instilled in the author a lifelong interest in elections and a healthy understanding of the importance of procedure." :)

Anyway, here's another excerpt, from pages 51-56, outlining several of the key problems with a direct popular-vote system:

A direct election system would strip away several all-important safeguards, inherent to the Electoral College system, that ensure a timely resolution and prevent a constitutional crisis—and all in the name of choosing the “right winner,” which is impossible to do with any certainty in a close election anyway, even if one accepts the dubious proposition that “the right winner” means nothing more than “the guy who got the most votes.”

One of the most important lost safeguards, and one that cannot be recaptured by any possible permutation or amendment of a direct election proposal, is the loss of the current system’s invaluable tendency to “quarantine” election disputes to individual states.  Even in the wildest imaginable Electoral College scenario, only a small handful of states would ever be truly “in play” and relevant to the determination of the winner.  In 2004, the only such state was Ohio; 2000, it was Florida; in 1960, Texas and Illinois; in 1888 and 1880, New York; in 1876, South Carolina, Louisiana, Florida and Oregon.  Outside of these states, any discrepancies or irregularities in those elections simply did not matter, either because they happened in a state too small to affect the national outcome or (more often) because the winning candidate’s margin in the state in question was too large to be seriously threatened.  In a direct election system, by contrast, each state’s individual margin does not matter; the state tallies only matter as part of a larger whole, so the difference between, say, a 1,000,000-vote lead in a given state and a 1,005,000-lead in that same state is of equal significance to the difference between a 2,500-vote lead and a 2,500-vote deficit.  In either case, the difference is 5,000 votes out of the national total.  This portends an enormous increase in the numbers of challenges and lawsuits in the event of a close election.  …

Moreover, because the challenges to a national popular election would originate in different state court systems, and probably the federal court system as well, it would be very difficult to streamline these lawsuits into a single case or small group of cases working their way up the appellate chain toward a single final resolution, as occurred in Florida in 2000.  Jurisdictional issues would likely prevent such streamlining, and even if not, consolidating cases from multiple states would create a nightmarishly complex set of facts for any court to analyze and adjudicate.  Simply put, the state-by-state electoral “quarantine” created by the Electoral College is the only thing holding back the floodgates of judicial chaos in the event of a close election.  A tight race under a direct election system would lead to a torrent of litigation that would make Florida 2000 look like an exercise in litigious restraint.  The end result could be paralysis and constitutional crisis, seriously undermining the system’s ability to fulfill the “undisputed and timely winner” purpose of presidential elections.

Another lost safeguard is the 200 years of history and precedent that underlie the Electoral College system and imbue it with legitimacy.  No matter how well-drafted a constitutional amendment creating a wholly new system might be, it is impossible to anticipate every conflict that could potentially arise.  Thus the courts will inevitably be drawn into adjudicating issues that have never been considered before, and while courts are routinely called upon to do that in other areas of the law, there ought to be a strong presumption against “reinventing the wheel” when it comes to the method of choosing the president of the United States, the most powerful official in our government.  The long history of the Electoral College creates public confidence that the system will be able to successfully navigate close elections, and that same history guides the courts and Congress as they consider how to follow procedures and resolve disputes.  We know what the rules are, we know how they have historically been interpreted, and we know they are not just going to change in mid-stream.  Essentially starting from scratch would create considerable uncertainty, and again, disrupt the certainty of an undisputed and timely winner. 

Yet another problem with a true direct election system is the elimination of the mid-December and early-January deadlines—which are, as noted earlier, more than mere deadlines, but are in fact beginning and ending dates of wholly independent stages in the election process.  This multi-stage aspect of the Electoral College system creates the aforementioned “instant replay” effect: once a new play starts, the old play can no longer be reviewed.  Once the electors have voted, and especially once Congress has counted and certified their votes, challenges to close elections in individual states become irrelevant, thus preserving the integrity of the result and preventing a legitimacy crisis if disputes arise later.  A true direct election system might have more malleable deadlines, as it would likely be a much more linear process.  Even after the “final” certification of the national vote tally is announced, it is not difficult to imagine it being revisited in the name of “counting every vote” if a late development threatens to change the result.  A state or even a federal court might adopt the logic of the Florida Supreme Court, which extended a statutory deadline on the basis of a legal interpretation that rested in part on its overarching belief that “the will of the people, not a hyper-technical reliance upon statutory provisions, should be our guiding principle in election cases.”  The point here is not to lambaste the Florida court or rail against judicial activism, but simply to point out that the battle cry “count every vote” is a compelling one, and well-meaning judges may be sorely tempted to follow its call even when important deadlines stand squarely in the way.  This may appeal to strict majoritarians, but the consequences for timeliness and certainty would be disastrous. 

A good example of what could occur in a fairly linear direct-election system, devoid of the sort of truly ironclad deadlines that characterize the Electoral College system, is the 2004 Washington gubernatorial race.  As noted earlier, vanquished Republican candidate Dino Rossi did not concede until he was defeated in a final court challenge on June 7, 2005—fully seven months after the election and five months after Gregoire’s inauguration.  Consequently, his election contest morphed into a “five-month legal effort to remove her” from office.  When the courts finally forced Rossi to give up, a spokeswoman for Gregoire—by this point roughly 10 percent done with her four-year term—said the decision “brings an end to a long campaign. This is a moment that will allow her and the state to move forward in addressing all of the pressing issues before us.”  It should go without saying that a similar scenario would be devastating on the national level.  The impact of such a cloud hanging of the president’s legitimacy would be far-reaching and crippling.  Thus, in an extremely close election, timeliness and certainty easily outweigh any “right winner” benefit, since the scenario in question could only occur in an extraordinarily close election, where the “right winner” is essentially a flip of a mathematical coin anyway.

Again, if you want to read the whole thing, just go to its SSRN page and download it.

Viva the Electoral College! :)

P.S. To those who volunteered to be "beta-testers" for the paper, I apologize for never sending you a copy. My plan was to make my "final" revisions earlier in the summer, then get input from the "beta-testers," make whatever changes might result from that, and then submit it to SSRN and the law reviews. But I didn't finish my revisions till much later than I hoped, and with deadlines approaching, I had to skip over the "beta" phase and pull the trigger on the submissions.

Comments

[Andrew]

Brendan, you know very well we're on the same page when it comes to the Electoral College. However, I have to wonder, does your paper actually explore, document, and/or quantify exactly how slipped deadlines and challenges to the president-elect's legitimacy would be "devastating", "far-reaching", and "crippling"? I have my own ideas why that might be the case, but I would hope for thoroughness' sake your paper covers those kinds of details in preparation for the potential majoritarian response to your wailing and gnashing of teeth, "So what?" Maybe sometime in the future I will have time to read such a lengthy paper but for now I'm hoping the author can give me a simple and quick answer. ;-)

[Alasdair]

Andrew - I have 'known' Brendan much less time than *you* have, and *I* do not belive it is even possible for said author to give a "simple and quick answer" to such a question !

(grin)

[Joe Loy]

Andrew, I think those Wailing & Gnashing sounds would emanate from the majoritarians pluralitarians, not from Brendan; and that their response would be less of a "So what?" than a "Uhh...WHAT?" :)

Alasdair, good point. A foolish Brevity is the Hobgoblin of small Mouths. Now back to the heather braes & haggis with ye. :>

Brendan: good work.

[Joe Loy]

PS: Brendan, also thanks for those Thanks. :)

(See, it's Like I always Say: "Proceduralists of the world, Unite! We have nothing to lose but our Audience." :)

[WizKid]

I believe I caught a typo in your abstract. Second line from bottom...

"Abstract:
It argues that, although imperfect, the Electoral College is superior to the alternatives and should be retained because it does an excellent job of balancing the two central purposes of presidential elections: choosing the right winner, and going(you probably meant doing) so in a reasonably timely and undisputed fashion."

The word "going" should be "doing"

[Brendan Loy]

D'oh! That was careless of me. Thanks, WizKid. You're quite right. I just fixed it.

Andrew, I do get into the importance of the office of the president a little bit when I discuss reasons why we need to avoid uncertainty about the president's identity, but I didn't talk too much about the specific conceptual reasons why legitimacy is important. I suppose I regarded those reasons as somewhat self-evident, but perhaps that was a mistake. It's something I might add in if I end up having an opportunity to revise the paper further (e.g., if I end up resubmitting in the spring).

[Andrew]

Joe, the sentence might have been long and convoluted, but the "So what?" was supposed to be from the majoritarians, not Brendan.

The key reason I poke on it is that, after the 2000 election, Democrats and liberals made quite a case of charging that Bush was not a legitimate president and doing everything in their power to undermine him. I wonder if we're getting to the point of vitriolic partisanship where any election that is closely contested leads to heated anger by the losing side, resorting to any and all sorts of legal and political challenges to the president's authority and legitimacy. If the EC doesn't save us from that problem now, the thinking might go, why should we listen to chicken-little arguments about how bad it'll be once we get rid of the EC? Alternatively, maybe some people would consider it a good thing to have a president neutered by repeated recounts and court challenges, with key positions going unfilled and the president unable to exercise any authority, de facto ceding power to lower bureaucratic levels, the states, and Congress.

[Joe Loy]

Andrew, I believe I got your meaning right, hence I wrote "...that their response" [i.e., the EC opponents'] "would be less of a 'So what?' than a 'Uhh...WHAT?'" :) ~ MY meaning being, that they (OK OK, Many of them, not All) will have No idea what Brendan is talking about :>. / Or, perhaps more descriptively, that they will rebut his Argument From Procedural Pickypickiness ;} with Theirs from Militant Ignorance, e.g., "Whaddaya MEAN, we can't accurately count all the Votes?? WTF??? Of COURSE we can! COUNT THE VOTES! COUNT THE VOTES! Just get rid of all these effing Halliburton Votingmachines!! FRAUUUD! FLOR-I-DUH! O-HI-YO!! LIE-BER-MAN!!!" / OKOK, so I just threw that last one in for Fun, so Sue me already :).

But as for your sentences being long & convoluted, Andrew, yeah, you wanna always be on guard against that. / WAW HAW HAW HAW hee hee Hoo-hah sometimes I crack me up ;}

"I wonder if we're getting to the point of vitriolic partisanship where any election that is closely contested leads to heated anger by the losing side, resorting to any and all sorts of legal and political challenges to the president's authority and legitimacy."

I don't think we're Getting To that point. I think we are At it. I fear we are stuck On it.

And to Think: it's all because poor Theresa LePore, in her honest & sincere effort to make her ballot more Readable (via larger typeface) to the Elderly folk of Palm Beach County, went Butterfly and thus (oops! :) accidentally boosted ould Paddy "Hang-the- Zionist-Lobby" Buchanan's Jewish vote to levels unDreampt-of in Yer Man's most fearsome nightmares :>.

[That's true y'know. Had only Patrick J.'s Palm Beach Irregulars correctly Punched out the votes they Intended to cast, the "Gore Wins Florida!" exitpoll projections would have proven Correct (by a Close but still Clear margin), there would have been no EC Inversion, no SCOTUS case, Al Gore would be finishing up his 2nd term, and global Warming, albeit not global Terrorism, would be Defeated today. ;]

[AnthonyH]

I like your excerpt, but I noticed one assumption that you make that I've seen made numerous times.

Your statements include "James Garfield defeated Winfield Hancock by just 9,070 votes, or 0.098 percent." The fallacious assumption is that the popular vote has anything to do with the election. It does not. It is definitely correlated, but it's only a correlation, not a direct measure. The candidates are not campaigning to win the popular vote.

I can guarantee that there were a lot of Democrats in Texas that didn't bother to vote during the 2000 election. Statewide offices were all going Red, and they knew that Texas' electoral college votes were going to be Red no matter whether they voted or not. In fact, there may have been plenty of Republicans who didn't vote for the same reason.

Knowing that their votes would be added to a national total would have changed the total number of voters.

I've never seen such a poll, but I'd like to see a poll question like this: "Have you ever skipped voting in a presidential election because you knew that your vote wouldn't affect the winner of the Electoral College votes for your state?"

Regardless of my statements above, I still have to agree with you that changing to a popular vote would increase the litigation level of Presidential races, and might cause problems like in Washington.

I like your idea of state-by-state quarantine, but I'd like to see some method where no state could be ignored on the campaign trail, yet we could keep the quarantine. Maybe some method by which we take the population of each state and determine a number of electors that would give an approximately equal population per elector. The number of electors would change with each election as population changed, and we could give some electors to Puerto Rico, D.C, Guam and other non-state areas (maybe combined, if the populations are not enough to get to the minimum). The candidate would get a percentage of the electors based on popular vote by state, and we would still have the quarantine we need.