Recounts: Elections in Overtime

Edward B. Foley

Public Law and Legal Theory Working

Paper Series
No. 209

September 18, 2012

This working paper series is co-sponsored by the

Center for Interdisciplinary Law and Policy Studies
at the Moritz College of Law

This paper can be downloaded without charge from the

Social Science Research Network Electronic Paper Collection:
[Link]

Electronic copy available at: [Link]

REVISED DRAFT: February 20, 2012
All rights reserved.

Recounts: Elections in Overtime

Edward B. Foley*

Introduction: Our National Impatience to Know Which Candidate Won

In the second half of the twentieth century, with the rise of television,

Americans developed an expectation that at the end of Election Day—after dinner

and after the polls closed—they could turn on their TV sets and watch the returns

come in. Doing so became something of a national ritual, akin to watching the

Super Bowl, particularly in presidential elections. Perhaps watching America’s

athletes in the summer Olympics would be a better analogy, since they occur on

the same quadrennial cycle as presidential elections.

Thus, when Americans tuned in on Election Night, they expected trusted

broadcasters, like Walter Cronkite or Tom Brokaw, to announce when each state

was declared for one of the competing presidential candidates. At some point into

the night—earlier some years, later in others—the announcer would proclaim that

one of the candidates had crossed the finish line by winning enough states to

achieve a majority of Electoral College votes. With that announcement, the

election would be over. The losing candidate would soon appear to make a

*
Director, Election Law @ Moritz, and Isadore & Ida Topper Professor of Constitutional Law,
Moritz College of Law at the Ohio State University.

Electronic
Electroniccopy
copyavailable
availableat:
at:[Link]
[Link]
concession speech. The winner would then make a victory speech, and

Americans would go to bed looking forward to the inauguration of the winner on

January 20.

As part of this ritual, however, the three major TV networks—CBS, NBC,

and ABC—would compete for viewers. (In this respect, election coverage differs

from sporting events, for which one network would have purchased exclusive

broadcast rights.) A major method of competition was for each network to

develop its own formula for “projecting” the outcome in each state, based on exit

polls of voters after they had cast their ballots. The networks would announce

these projections on Election Night before the government authorities in the states

were finished reporting the actual returns from all the polling locations statewide.

Given this practice of making projections, the networks—and thus their

viewers—began to distinguish between the entirely unofficial status of their own

projections and the official status of the numbers reported on Election Night by

government authorities. “With 60% of precincts reporting, Candidate Smith is

leading Candidate Jones by 30,000 votes. Our network, however, is not yet

prepared to project Smith the winner, as our analysis shows that the remaining

40% of precincts are in places where Jones might be able to make up the

difference.” This terminology would cause viewers to think that once 100% of

the precincts had reported, the election would be officially over in the state.

Electronic
Electroniccopy
copyavailable
availableat:
at:[Link]
[Link]
 On the contrary, the official vote-counting process is just beginning with

the report of precinct returns. At a minimum, the arithmetic tabulation of those

returns needs to be verified as part of the “canvassing” of those returns. Absentee

ballots that arrived too late to be counted as part of those initial returns, but are

still eligible to be counted if postmarked on time, still need to be counted.

Depending on the particular rules in a state, an audit or partial recount of ballots

needs to be conducted in order to assure the accuracy of the vote-counting

machines. These steps and others must be conducted before the result of an

election can be certified. An election is never officially over until this

certification occurs, which may take place days or weeks after the ballots were

cast on Election Day.

The average American viewer watching the election returns in 1972 or

1992 most likely would have been unaware of all this. The expectation was that

elections were over on Election Night, or early the next morning if it took an extra

long time for all the official returns to come in. The networks might “jump the

gun” by a few hours or so with their projections, but the average TV viewer

thought that the election was officially over once the ballots in all the precincts

had been counted and these results announced. After all, the losing candidate

always conceded by late that night or early the day. Therefore, the election must

be over.

Electronic
Electroniccopy
copyavailable
availableat:
at:[Link]
[Link]
 Thus, Americans were psychologically unprepared for what happened in

2000. They were expecting to know who won that year’s election by the end of

Election Night or early the next day. It did not help, of course, that the networks

changed their projections during the course of the night, or that the Democratic

candidate, Vice-President Al Gore, called his Republican opponent, Governor

George Bush of Texas, on the telephone to concede, only to call Bush back an

hour later to retract his concession. But even if none of that had occurred, the

nation would have been unsettled—in other words, unprepared psychologically to

go through the official process of canvassing the returns without knowing which

candidate actually won until after the certification of the canvass. The idea that

the election was close enough that we just might not know the result for a week or

two, while verifying the accuracy of the returns and counting late-arriving but

still-eligible absentee ballots, just seemed too much for the American psyche to

take.1 Indeed, even as Americans lived through this experience, and the resolved

presidential election of 2000 continued into its fourth and fifth week, the country

did not seem to take heart that it could survive this kind of uncertainty. Instead,

nerves became increasingly frayed, and the U.S. Supreme Court evidently sensed

1
On November 12, 2000, five days after the casting of ballots in that year’s presidential election,
the celebrated New York Times columnist R.W. Apple wrote a piece with the subtitle “The Limits
of Patience.” The column opened with these lines: “Another week and no more. By next
weekend, a group of scholars and senior politicians interviewed this weekend agreed, the
presidential race of 2000 must be resolved, without recourse to the courts.” R.W. Apple, Bush
Sues to Halt Recount in Florida; The Limits of Patience, N.Y. TIMES, Nov. 12, 2000, at p. 1.

Electronic copy available at: [Link]

that it was reflecting the prevailing national mood by bringing the uncertainty to

an end.

Recounts, Both Routine and Exceptional

This national need to know who won the presidential election immediately

after the polls have closed, if it indeed still exists after our experience in 2000, is

at odds with our attitude about other elections. Inevitably, in each general

election, there is some race somewhere in the country that remains undecided for

several weeks or maybe even many months. It could be a city council race, or a

judicial election for a seat on a local court, or maybe even an election for a

member of the state legislature or the U.S. House of Representatives.

The idea that such a race could be close, to be decided by just a handful of

votes, is not alarming or bizarre. Nor is the idea that the legal system would have

mechanisms to make sure, before a winner is officially and finally declared, that

all the votes have been counted accurately and, in some cases, recounted if

necessary. Average citizens accept that these mechanisms may take time, and

thus knowing the outcome of the election may be delayed. They even accept that,

in some circumstances, there may be credible allegations of mistakes or

improprieties in the vote-counting process that would require the involvement of a

court to consider the evidence that might either support or refute such claims.

Electronic copy available at: [Link]

Thus, if they gave the matter much thought, they would recognize—and

tolerate—the fact that this judicial involvement would add some significant

additional delay before a winner of the election could be conclusively declared.

Indeed, these sorts of mechanisms and judicial procedures exist in every

state in the nation and are routinely deployed to eliminate uncertainties about

which candidate won a particularly close race. Little media attention is given to

these recounts and judicial proceedings to settle an election’s outcome. They

almost always involve local races where the stakes are not especially high. The

existing procedures usually work reasonably well to achieve closure, so that one

of the candidates can take office for the remainder of the term, and then when that

term is over, the voters will have another chance to express their will about who

should get to hold that office.

But the circumstances are very different when the election concerns not a

low-stakes local race, but instead a high-stakes statewide office, like governor.

First of all, in each state, it is extremely rare to have a major statewide election

unsettled after Election Night. The mathematical “law” of large numbers

accounts for this statistical fact. While a small local election involving only a

thousand ballots could easily end up with only a 10-vote margin between the two

leading candidates—and thus the necessity of a recount and perhaps a judicial

lawsuit to determine whether that 10-vote margin holds up as valid—a large

statewide election involving a million ballots or more is much less likely to end

Electronic copy available at: [Link]

up in such an easily disputable 10-vote margin. Think of it this way: say two

candidates run neck-and-neck in a small local election with only one thousand

voters, and they split the vote almost evenly 50.1% to 49.9%; the outcome is a 2-

vote margin. But if two candidates receive the same neck-and-neck percentages

in a large statewide race involving one million ballots, the result is a 2,000-vote

margin, which is not as easily contestable.2 Thus, states almost never experience

a disputed gubernatorial election, and thus neither their citizens nor their public

officials are prepared for such a dispute in the freakish circumstance that it

actually happens to them. Like presidential elections, gubernatorial elections

virtually always end with a concession speech on Election Night, and when that

does not happen a certain degree of anxiety in the public mood begins to develop.

Moreover, there is a lot of political power in the hands of the governor. It

really matters which candidate gets to hold that office. Therefore, if the

extraordinarily unlikely event does occur and initial returns show a gubernatorial

election to be within 100 votes or so, it is a prize that seems tantalizingly within

reach and very much worth fighting for. Thus, almost immediately the two

competing candidates and their campaigns begin to put intense pressure on the

state’s legal apparatus for recounting ballots and resolving related vote-counting

disputes.

2
In a study of all statewide elections from 2000 to 2009 in the U.S., which totaled 2884 races, the
organization FairVote found that only eighteen ended up in a recount, and of these only three
resulted in a reversal of the original outcome. See Rob Richie & Emily Hellman

Electronic copy available at: [Link]

 Yet the state’s legal apparatus is inevitably rusty, at least with respect to

handling a statewide recount. For the reasons already indicated, there probably

has not been another statewide recount for decades. No one in state government

is familiar with how the recount should proceed, and yet everyone in state

government has an interest in—or at least a predisposition towards—which

candidate for governor they would like to prevail in this newly emerging dispute

over who got more votes. The combined attributes of inexperience and partiality

among the public officials who are supposed to supervise the recount and resolve

the uncertainty over the outcome is a recipe for significant controversy.

Indeed, among the relatively small number of disputed gubernatorial

elections in the nation’s history, some have not gone well at all. For example, in

1879, even after all the bloodshed of the national Civil War, Maine was on the

brink of its own state-specific civil war when the incumbent Democratic governor

and his supporters manipulated the canvassing of local returns so he could stay in

office despite the apparent defeat at the polls of his reelection bid. Only the

heroic and patient intervention of a former Civil War hero, Joshua Chamberlain,

who narrowly escaped an assassination attempt on his life, managed to convince

the Democrat incumbent and his partisans eventually to step aside without the

necessity of shots being fired.3 A similarly precipitous conflict had occurred over

Pennsylvania’s gubernatorial election of 1838, known there as the Buckshot War,

3
For an account of Chamberlain’s steady leadership during Maine’s moment of crisis, see
Chapers 8 & 9 of John J. Pullen, JOSHUA CHAMBERLAIN: A HERO’S LIFE AND LEGACY (1999).

Electronic copy available at: [Link]

because of the armed forces arrayed against each other after the incumbent party,

the Whigs, attempted to hold on to power by manipulating the returns from

polling places in Philadelphia. At the height of that conflict, there was an attempt

to kidnap—and perhaps assassinate—Thaddeus Stevens, who was helping to

orchestrate the Whigs’ efforts to keep power (and who would go on to become a

leading Radical Republican during Reconstruction after the Civil War). Stevens

had to jump out of a window and run for his life, in order to escape his assailants.4

William Goebel, the Democratic candidate for governor of Kentucky in

1899, was not so fortunate. He was killed by an assassin’s bullet during the huge

dispute that engulfed the state over the outcome of that election. The dispute

went all the way to the U.S. Supreme Court, but unlike a century later in Bush v.

Gore (2000) the Court in Taylor v. Beckham (1900) explicitly refused to let a

claim that the wrongful counting of votes violated the Fourteenth Amendment of

the federal Constitution serve as a basis for the nation’s highest Court to become

involved in the vote-counting dispute. Consequently, Goebel’s running mate had

to serve the term that Goebel, after the Court’s refusal to intervene, conclusively

was declared to have won.5

4
Thaddeus Stevens’ role in the Buckshot War is depicted, perhaps a bit more favorably than in
other accounts, in Hans L. Trefousse, THADDEUS STEVENS: NINETEENTH CENTURY EGALITARIAN
57-60 (1997).
5
The sordid story of this dispute is told in Tracy Campbell, DELIVER THE VOTE: A HISTORY OF
ELECTION FRAUD, AN AMERICAN POLITICAL TRADITION—1742-2004, at 106-110 (2006). The
citation for the U.S. Supreme Court’s decision is Taylor v. Beckham, 178 U.S. 548 (1900).

Electronic copy available at: [Link]

 Fortunately, not all close gubernatorial elections in the nation’s history go

as badly as the ones just described. In 1839, one year after the awful Buckshot

War in Pennsylvania, the Democratic candidate for governor of Massachusetts

won the election by just a single vote. Yet even though the Whigs held power

there, as they had in Pennsylvania, there was no imminent risk of violence in

Massachusetts. The incumbent Whig governor, Edward Everett, honorably

quashed any sentiment of manipulating the canvass in order to keep him in power

when an honest count showed otherwise, even by the literally slimmest of

margins!6

In 1855, the incumbent Wisconsin governor was not as honorable as

Everett and, like the Pennsylvania and Maine incumbents in 1838 and 1879,

attempted to use the power of his office to manipulate returns in order to cling to

power. But the Wisconsin Supreme Court issued a strong decision asserting its

judicial power to make sure the votes were counted legally and honestly, and the

incumbent governor backed down before the dispute there escalated to anything

like the dire crises that occurred in Pennsylvania and Maine. The peaceful

resolution of this Wisconsin dispute was a major victory of the rule of law to

6
At the time, it was necessary under Massachusetts law for a gubernatorial candidate to win a
majority, and not just a plurality, of the votes cast in order to win the election outright; otherwise,
the state’s legislature was entitled to choose which candidate it preferred. Edwards’s Democratic
opponent won the exact number necessary to make a majority; one less vote would have put the
election in the hands of the legislature, which would have chosen Edwards since both houses were
in the hands of his party, the Whigs. The details of this election, including Edwards’s
magnanimity can be found in Paul Revere Frothingham, EDWARD EVERETT: ORATOR AND
STATESMAN 150-155 (1925).

10

Electronic copy available at: [Link]

protect the operation of democracy in the specific context of a fight over the

counting of ballots to determine who will hold the highest office in the state.7

A century later, Wisconsin’s neighbor Minnesota was able to employ its

judiciary in an innovative way to conduct a remarkably successful recount of an

exceptionally close gubernatorial race. In 1962, initial returns showed the

incumbent governor winning by just 58 votes out of over 1.2 million cast. After

some initial procedural skirmishing, which threatened to leave the Justices of the

Minnesota Supreme Court looking like they were lining up on either side of the

fight depending on each of their own partisan affiliations, the Chief Justice

decided that the recount should be conducted by a special three-judge court whose

members were chosen by equal agreement of the two sides. The Chief Justice’s

move was not unlike convening an arbitration panel to resolve a dispute between

management and labor, where each side must agree upon the arbiters. In any

event, the Chief Justice’s gambit worked. Lawyers for the two candidates picked

three judges for the special recount court: the key was their agreement on one of

these judges as being genuinely independent, while each of the other two judges

were associated with each of the two major political parties in the state. The

recount conducted by this special court showed, in fact, that the challenger had

won by 91 votes. The incumbent then graciously accepted this defeat. Not only

7
See Chapter 9 of John Bradley Winslow, THE STORY OF A GREAT COURT (1912) (a history of the
Wisconsin Supreme Court).

11

Electronic copy available at: [Link]

was the entire process peaceful, but it was also acknowledged by the losing side

to be as fair as feasible.8

Our two disputed presidential elections (and some “near misses”)

The United States has experienced only two presidential elections affected

by serious vote-counting controversies: 1876 and 2000. They are thus even rarer

than seriously disputed gubernatorial elections. The reason is that, given the way

the Electoral College works, the vote for President must be extremely close in a

particular state and that state’s Electoral College votes must be crucial to winning

an Electoral College majority (as required by the Constitution).

There have been years where the popular vote for President in a particular

state has been mired in controversy, but that state’s Electoral College vote did not

make a difference in the winning candidate’s ability to achieve an Electoral

College majority. For example, in 1872, as a kind of harbinger of what would

become consequential four years later, Congress could not figure out which

candidate had won the Electoral Votes of Louisiana and Arkansas. But Congress

decided not to count the Electoral Votes from those two states at all, because

President Grant easily won reelection that year without them. Similarly, in 1960,

Hawaii’s initial returns showed Richard Nixon winning that state, whereas a
8
The definitive account of Minnesota’s successful resolution of its disputed 1962 gubernatorial
election is Ronald F. Stinnett & Charles H. Beckstrom, RECOUNT (1964).

12

Electronic copy available at: [Link]

subsequent recount had the state going for John Kennedy. There was a legitimate

question whether the result of the recount had occurred soon enough under the

U.S. Constitution for Congress to count the state’s Electoral Votes as going to

Kennedy rather than Nixon. But Congress did not bother to render a formal

verdict on this constitutional question since Kennedy did not need Hawaii to

achieve an Electoral College majority.9

Apart from 1876 and 2000, there have been a few other presidential

elections where initial Election Night returns indicated the possibility of a lengthy

battle over which candidate could obtain an Electoral College majority, but in

each of these other instances a protracted outcome-determinative ballot-counting

controversy was avoided. In 1884, whichever candidate won New York would

win the White House, and the outcome in New York remained unsettled for a

couple of weeks, while local canvassing boards reviewed their returns under the

watchful eyes of representatives of both parties. When the canvass was done,

however, Republicans were satisfied that Grover Cleveland’s 1047-vote victory in

the state was valid, and they did not try to contest it any further. Similarly, in

1916, Woodrow Wilson’s reelection bid depended upon the outcome of

California, and that state’s results also remained unsettled for a couple of weeks

that year. Again, however, after scrutinizing the local returns, Republicans

9
The 1872 and 1960 situations are discussed in Nathan L. Colvin & Edward B. Foley, The Twelfth
Amendment: A Constitutional Ticking Time Bomb, 64 U. MIAMI L. REV. 475 (2010).

13

Electronic copy available at: [Link]

satisfied themselves that Wilson’s 3,420-vote victory in the state could not (and

should not) be challenged as an erroneous result.10

Both 1884 and 1916 were before the television era, during a time when the

country was apparently untroubled by a presidential election taking two weeks or

so to become resolved. 1960 was in the early years of television, and its

avoidance of a dispute that had the potential of becoming protracted may have

helped develop the expectation, well-settled by 2000, that the nation should know

the winner of a presidential election soon after the polls close on Election Day.

Nixon’s decision to quickly concede, when he might have decided to fight for a

while, prevented the kind of intense scrutiny of the nation’s ballot-counting

procedures that occurred forty years later.

Although Hawaii was never a factor in determining the Electoral College

winner in 1960, both Illinois and Texas were. Based on initial returns from

around the country, Kennedy could achieve an Electoral College majority with

either state; Nixon needed both to win. Republicans had reasons to be suspicious

of the returns from both states. There were rumors of Mayor Daley’s Chicago

machine stealing Illinois for Kennedy. Likewise, Kennedy’s running-mate

Lyndon Johnson had gotten to the U.S. Senate as the result of a 1948 election

tainted by falsified election returns,11 and Nixon supporters feared that Johnson’s

10
These vote margins can be found in CQ PRESS, PRESIDENTIAL ELECTIONS 1789-2008 (2009).
11
Robert Caro’s riveting narrative of Johnson’s 1948 victory is told in THE MEANS OF ASCENT,
the second volume of Caro’s prize-winning THE YEARS OF LYNDON JOHNSON.

14

Electronic copy available at: [Link]

political friends back in Texas might have done something similar for the

Democratic ticket in 1960. But as Nixon assessed the situation, the odds of

overturning the initial returns in both Illinois and Texas seemed too much of a

long-shot. Even if he could overturn the narrower margin in Illinois of 8,858

votes, the much larger margin in Texas—46,257—seemed insurmountable,

especially considering that Texas’s legal machinery at the time was deemed

incapable of conducting any kind of impartial statewide recount. (Texas, one

must recall, was essentially a one-party state, with Democrats in complete

control.) Nixon wanted to give himself the possibility of a political comeback,

which in fact would occur in 1968, and he thought he would ruin any chance of

that if he fought results in Illinois and Texas and did not prevail. Consequently,

shortly before 10 a.m. California time on the morning after Election Day, Nixon

publicly conceded defeat, thereby short-circuiting the possibility of any publicly

transparent and rigorous recount of the presidential votes in those two outcome-

determinative states.12

Thus, 1876 and 2000 remain (at least so far) the only two presidential

elections where there developed serious ballot-counting disputes, which caused

the outcome to remain unresolved for more than a couple of weeks. These two

disputes, despite the historical distance of 124 years between them, share some

12
Two new histories, timed to commemorate the fiftieth anniversary of the 1960 election, review
Nixon’s decision to concede quickly: W.J. Rorabaugh, THE REAL MAKING OF THE PRESIDENT:
KENNEDY, NIXON AND THE 1960 ELECTION (2009), and Edmund F. Kallina, KENNEDY V. NIXON:
THE PRESIDENTIAL ELECTION OF 1960 (2011).

15

Electronic copy available at: [Link]

similarities. Florida, for example, played a key role in both. But there are also

important differences between the two episodes. The 1876 dispute involved

multiple states, not just one, and Congress resolved the 1876 with the help of a

specially created Electoral Commission, whereas the 2000 dispute ended after the

U.S. Supreme Court order a halt to the recounting of ballots that was then

underway in Florida.13

1876

The presidential election of 1876 occurred as the South was increasing its

resistance to Reconstruction. The Republicans were the party of Reconstruction,

while the Democrats were the party of the resistance. The Civil War hero,

Ulysses S. Grant, the two-term Republican incumbent in the White House, was

not running again, and the Republicans nominated another Civil War general,

13
Although obviously a product of its era, and thus not reflecting a contemporary understanding of
racial equality, the most comprehensive single book on the political battle to prevail in the
controversy over counting votes in the 1876 election—and one that is fair in its description of both
the Democrats and Republicans involved in the fight—is Paul Haworth, THE HAYES-TILDEN
DISPUTED PRESIDENTIAL ELECTION OF 1876 (1906). For more modern accounts, written in the
aftermath of 2000, William H. Rehnquist, CENTENNIAL CRISIS: THE DISPUTED ELECTION OF 1876
(2005) (written by the Chief Justice of the United States during Bush v. Gore), and Michael F.
Holt, BY ONE VOTE: THE DISPUTED PRESIDENTIAL ELECTION OF 1876 (2011) (an incisive analysis
by an eminent contemporary historian, but ultimately somewhat more selective in its coverage of
the post-voting dispute than Haworth).
The literature on Bush v. Gore and the presidential election of 2000 is voluminous. For a
journalistic approach that reads like a “page-turner” novel, although told largely from the
perspective of the Gore campaign, see Jeffrey Toobin, TOO CLOSE TO CALL: THE THIRTY-SIX DAY
BATTLE TO DECIDE THE 2000 ELECTION (2002). If one prefers a movie version of this story, even
more favorable to the lawyers who fought to count more votes for Gore, there is HBO’s RECOUNT
(2008) (available on DVD). A more academic study is Charles L. Zelden, BUSH V. GORE:
EXPOSING THE HIDDEN CRISIS IN AMERICAN DEMOCRACY (2010), although it arguably also tilts
towards the pro-Gore point of view. A version of the story from the other side is Robert Zelnick,
WINNING FLORIDA: HOW THE BUSH TEAM FOUGHT THE BATTLE (2001). A useful primer on the
legal issues involved is Abner Greene, UNDERSTANDING THE 2000 ELECTION: A GUIDE TO THE
LEGAL BATTLES THAT DECIDED THE PRESIDENCY (2001).

16

Electronic copy available at: [Link]

Rutherford B. Hayes of Ohio, as their presidential candidate. The Democrats

chose a northerner, Samuel J. Tilden of New York, who had earned a reputation

as a reformer fighting political corruption.

Initial returns on Election Night showed Tilden winning enough states to

be just one Electoral Vote shy of a majority, with the outcome uncertain in three

southern States: Florida, Louisiana, and South Carolina. Reconstructionist

Republicans still controlled the vote-counting process in those three states, and

word went out from national party leaders that the local Republicans should use

their political power to make sure that Hayes, not Tilden, was declared the winner

all three of those states. Hayes needed all three to eke out an Electoral College

victory of 185-184, whereas Tilden needed just one of the three to put him over

the top.

Sure enough, the Republicans did use their control of the canvassing

boards in all three states to derive final vote totals that favored Hayes over Tilden.

Democrats, both nationally and locally, screamed of the fraud that the

Republicans had perpetrated. Republicans retorted, with justification, that in all

three states Democrats had committed horrendous acts of voter intimidation

aimed at preventing black citizens from exercising their right to vote, now

guaranteed by the Fifteenth Amendment to the U.S. Constitution as part of

Reconstruction. Historians have generally concluded that, if the vote-casting

process in these three states had been fair, then Hayes would have won the three

17

Electronic copy available at: [Link]

states with the support of black voters who had been unconstitutionally

disenfranchised. Yet, if one asks the question whether the vote-counting was fair,

it seems as if at least in Florida Republicans improperly manipulated the count of

the lawful ballots that had actually been cast, to deprive Tilden of a victory that he

would have received according to an honest count of those ballots.

In any event, the Democrats were not about to acquiesce in the vote-

counting fraud that they believed had been perpetrated against them. In each of

the three states, when it came time for the Electoral College to meet and cast the

state’s Electoral Votes, the Democrats had their slate of presidential electors hold

a separate meeting on the ground that they believed that their electors were the

ones who were entitled to cast the state’s votes for president. The problem for the

Democrats, however, was that the authority for their Electoral College meetings

was dubious at best. In each of the three states, on the date that the Republican

electors separately met and voted for Hayes, the Republican governor of that state

certified that these Republican electors had been duly appointed according to the

results as canvassed by the state’s official canvassing board. The best that the

Democrats could do was in Florida, where the state’s Attorney General—a

Democrat—certified that the Democrat electors were the valid ones according to a

proper count of the ballots, and that the governor’s contrary certification was

based on the fraudulent conduct of the state’s canvassing board. Later, the

Democrats in Florida attempted to bolster their case, by winning a lawsuit in state

18

Electronic copy available at: [Link]

court that their electors were the rightful ones and by then having the state’s

newly inaugurated governor—another Democrat—issue a new certification that

the pro-Tilden electors were indeed the state’s properly chosen ones. But these

new developments came long after the date on which, under the U.S. Constitution,

the electors were required to have met and voted for president. (As the

Democrats’ position in Florida was stronger than in Louisiana or South Carolina,

there is no need to consider the details of these two other southern states.)

Meanwhile, the Democrats thought they had figured out a trick that would

guarantee them the one additional Electoral Vote they needed for the majority that

would place Tilden in the White House. Although Hayes had undoubtedly won

the popular vote in Oregon, there emerged the question of whether one of the

Republican electors in the state was eligible to serve in that perfunctory role. This

particular elector was an employee of the U.S. postal service, an arm of the

federal government, but the U.S. Constitution said that “no Person holding an

Office of Trust or Profit under the United States, shall be appointed an Elector.”

Using an available state law, Republicans in Oregon had a way around this

problem: this particular elector could resign both his postal position and his status

as an elector, and the remaining valid electors could fill the vacancy by

reappointing the same person now that the cloud of his federal job had been

removed. But the governor of Oregon was a Democrat, and this fact caused

national Democrats to believe they held a winning hand. The Oregon governor

19

Electronic copy available at: [Link]

certified that, because the Republican elector was disqualified at the time the

citizens voted for him, the Democratic elector with the highest number of votes

was entitled to serve in his place. On the required day for the electors to meet,

this Democrat elector purported to meet with substitutes for the other Republican

electors, and he cast his one Electoral Vote for Tilden while the others cast their

votes for Hayes.

The Republicans in Oregon would have none of this, and they held their

own separate Electoral College meeting, and the Secretary of State was willing to

certify that they were entitled to do so, because they had all won the most votes in

the state (and the single problematic Republican elector had been reappointed

after resigning his positions, as permitted by state law). The Democrats thought

they had the situation beat: if the governor’s certificates from Florida, Louisiana,

and South Carolina would give those southern states to Hayes, then the Oregon

governor’s certificate would surely give them the one necessary Electoral Vote

for a Tilden victory. Conversely, if the Oregon governor’s certificate was not

good enough, because underneath it the Republican elector had really won the

state’s popular vote, then this argument would surely mean the Democrats would

prevail in the dispute over Florida, since underneath the first governor’s certificate

Tilden’s electors had really won the popular vote there. Either way, the

Democrats thought they could not be prevented from reaching an Electoral

College majority.

20

Electronic copy available at: [Link]

 Thus was the situation in January of 1877, as the partisan fighting over

who won the Hayes-Tilden election moved from the states to Washington, D.C.

At the time, there were no rules for what should happen next. The U.S.

Constitution stated only: “The President of the Senate shall, in the presence of the

Senate and the House of Representatives, open all the certificates and the votes

shall then be counted.” Which certificates? Multiple and conflicting certificates

had been sent from the three southern states and Oregon. Counted by whom?

Who would decide which certificates from these states would be counted? The

Constitution, using the dreaded passive voice, left this most crucial question

unanswered, and Congress had not enacted a law ahead of this election to fill in

this constitutional gap.

When this dispute reached Washington, Republicans controlled the Senate

and Democrats the House of Representatives. Thus, there was a risk of a serious

deadlock over how to resolve the dispute. Moderates from both parties in both

chambers forged a temporary compromise: Congress would create a special

Electoral Commission with five Senators, five Representatives, and five Justices

of the U.S. Supreme Court to consider each of the states that had sent multiple

certificates of Electoral Votes. The Commission’s decision with respect to each

state would prevail unless overruled by both the Senate and the House.

The Democrats thought that they had succeeded once they had secured

this compromise. By design, the Commission was supposed to be balanced

21

Electronic copy available at: [Link]

between the two parties: the ten congressional members were split between five

Republicans and five Democrats. Even four of the five Justices were understood

to be balanced between two Democrats and two Republicans. The fifth Justice

was supposed to David Davis, a Lincoln appointee whom the Democrats

perceived as genuinely independent (if not perhaps a little favorable to their side).

But Davis refused to serve on the Electoral Commission, having just been

appointed to the U.S. Senate by the Illinois legislature. In his place, Justice

Joseph Bradley was chosen as the fifteenth member of the Commission, but he

was another Republican.

By strictly partisan 8-7 votes in each case, the Electoral Commission ruled

in favor of counting the Electoral Votes cast by the Republican electors (and thus

rejecting the Electoral Votes of the Democratic electors). Justice Bradley

explained how he could rule for the Republicans electors in Florida but not the

one Democratic elector in Oregon. It was not the governor’s certificate that was

dispositive, but instead the laws of the state in place on the date that the electors

were required to cast their Electoral Votes under the U.S. Constitution. In

Florida, as of that date, state law had given its canvassing board the power to

declare the winner of the state’s popular vote. Even if that declaration was

inaccurate, or worse fraudulent, the canvassing board’s position was what

prevailed according to state law. Bradley acknowledged that a state’s laws could

permit a canvassing board’s inaccurate or fraudulent declaration to be overturned

22

Electronic copy available at: [Link]

by a state court, as in fact happened in Florida, but—and this was the crucial point

for Bradley—any such judicial ruling needed to happen by the date on which the

electors were constitutionally required to meet. Because the canvassing board’s

declaration that the Hayes electors had won the popular vote still stood as the

state’s official pronouncement as of the day that the electors met, in Bradley’s

opinion this pronouncement must be accepted as authoritative and binding in the

federal counting of the state’s Electoral Votes. With respect to Oregon,

conversely, it was clear that state law permitted the reappointment of the

Republican elector who had resigned because of his post office job, and thus for

Justice Bradley the Democratic governor’s contrary certification carried no

weight.

Whether or not Justice Bradley’s reasoning was analytically sound and

constitutionally justified, Democrats viewed it as self-contradictory and motivated

either by partisanship or some worse form of corruption. (There were allegations,

never proved and most probably without any foundation, that Bradley accepted

Republican railroad money in exchange for his rulings on the Commission.) For a

while, some recalcitrant Democrats considered trying to repudiate the

Commission’s work by asserting the constitutional prerogative of the U.S. House

of Representatives to elect the President itself directly whenever no candidate

wins a majority of Electoral Votes. But in the end enough Democrats acquiesced

to permit, based on the Commission’s rulings, the President of the Senate to

23

Electronic copy available at: [Link]

declare that Hayes had won an Electoral College majority. It helped southern

Democrats accept this defeat that close confidants of Hayes pledged that he would

withdraw federal troops from the South if and when he was inaugurated President.

Once in power, Hayes made good on this pledge, and thus the processes of

political compromise that ultimately paved the way for his peaceful entry into the

White House brought an end to Reconstruction.

2000

A decade after the Hayes-Tilden dispute, Congress passed the Electoral

Count Act of 1887 in the event that another disputed presidential election might

occur. It took ten years to pass the statute because Congress remained divided

between the two parties during that period, and neither party wanted to cement

into law a vote-counting procedure that might deprive it of an advantage it might

obtain if and when the next disputed presidential election arose. But 1884 had

been another exceedingly close presidential election (as mentioned above), and

Congress did not want to head into 1888, which had all the signs of being very

close again (and was, but as close as 1884), without a statutory mechanism for

dealing with a dispute like the one over the 1876 election.

The Electoral Count Act of 1887 is an extremely convoluted and

confusing statute, the interlocking clauses of which almost defy human

24

Electronic copy available at: [Link]

comprehension, but apparently it was the best that Congress could do under the

circumstances. Essentially, and necessarily oversimplifying somewhat, the Act

replaces the Electoral Commission that was used to break the Hayes-Tilden

deadlock with a procedure by which the governor’s certification of a state’s

Electoral Votes automatically prevails in the event that the Senate and House of

Representatives are divided on which of two or more competing certificates

represents the state’s authentic Electoral Votes. But the statute also contains a

provision, known as the Safe-Harbor section, which purports to bind both houses

of Congress to the state’s own resolution of a dispute over a state’s Electoral

Votes if both the state’s resolution is pursuant to procedures that were in place

before citizens cast their ballots for presidential electors and the resolution of the

dispute is complete six days before the day on which the state’s presidential

electors are constitutionally required to meet. Thus, if one of two certificates of

Electoral Votes from a state complies with this Safe-Harbor provision, under the

terms of the statute both houses of Congress are supposed to accept it (and thus

reject the other certificate). Presumably, however, if the two houses of Congress

act politically, and thus each simply votes for the certificate that supports the

presidential candidate of its own party, then the statute seems to say that the

governor’s certificate prevails, even if one house of Congress was acting in direct

defiance of its statutory obligation to honor the certificate that complied with the

25

Electronic copy available at: [Link]

Safe-Harbor provision. In this respect (among other problems), the statute is

somewhat at odds with itself.

The intricacies and potential inconsistencies of the Electoral Count Act

never mattered until 2000. As the dispute over the ballots from Florida unfolded

that year, observers began to wonder what would happen if the dispute reached

Congress. The prospects did not look pretty. Republicans would control the

House of Representatives, with the Senate split 50-50, giving Vice-President (and

presidential candidate) Al Gore the power to break a tie vote in the Senate.

Would he break a tie in favor of himself? If so, and assuming the Senate and

House then differed over which of two certificates from Florida to count, the

Electoral Count Act seemed to indicate that the certificate bearing the governor’s

signature should prevail. The governor of Florida happened to be Jeb Bush, the

brother of Gore’s opponent George W. Bush.

Perhaps the inadequacies of the Electoral Count Act, at least as it applied

the situation unfolding in 2000, is what caused the U.S. Supreme Court to become

involved in the dispute. Many have so speculated, although the Court has never

explained its reasons for taking the case except to say that it raised important

questions of constitutional law in a setting of obvious national significance. In

any event, Al Gore’s concession speech the day after the U.S. Supreme Court

decision in Bush v. Gore ended the dispute and thus avoided the necessity of

26

Electronic copy available at: [Link]

Congress having to struggle with applying the archaic and untested provisions of

the Electoral Count Act.

But for the U.S. Supreme Court to play the decisive role that it did, the

case had to come to it. (The Court has no power to reach out and grab cases that

it wishes to decide even if no one files the case in the Court.) To understand then

how the disputed presidential election of 2000 got to the U.S. Supreme Court, one

must follow its journey from November 7, Election Day, to December 12, the day

of the Court’s decision in Bush v. Gore.

After Gore retracted his concession to Bush at about 2:30 a.m. (central

time) during the night after Election Day, Gore’s lawyers began to look for ways

they could overcome the roughly 1600-vote deficit that then existed in the initial

returns. For a statewide election that close, Florida required an immediate

machine recount, meaning that local election officials were required to re-feed

paper ballots into their vote-counting machines. This procedure lowered Bush’s

lead to only 229 votes, but Gore was still behind.

Gore’s lawyers considered challenging the so-called “butterfly ballot,”

which undoubtedly had confused some elderly voters who, wishing to vote for

Gore, accidently voted instead for Pat Buchanan, the ultraconservative

commentator who was running as a candidate of the Reform Party founded by

Ross Perot. But Gore’s lawyers concluded that there was no legal remedy

available for the “butterfly ballot” problem. These ballots, after all, had been cast

27

Electronic copy available at: [Link]

for Buchanan even if the voter intended otherwise, and a Democratic official had

specifically approved the “butterfly ballot” for use in the election. Moreover,

there was no way to hold a “do over” just for the voters adversely affected by the

“butterfly ballot.”

Consequently, Gore’s lawyers focused their attention on the separate

problem of so-called “hanging chads,” which referred to partially dislodged bits

of paper dangling from punch-card ballots. In each instance of a hanging chad,

the voter had attempted to indicate their choice by removing the chad from its

perforated box on the ballot, but had failed to remove the chad completely.

Voting machines often could not read the vote intended by the voter, because the

hanging chad blocked the light from going through the hole in the punch-card

ballot, which was how the machine recognized a vote on this kind of ballot.

Gore’s lawyers wanted to get local officials in Florida to count by hand

votes that the machines may have missed because of the obstruction caused by the

hanging chads. This strategy presupposed that ballots with hanging chads could

count as valid votes even if not recognizable by the machines. But if it was the

duty of the voter to completely dislodge a chad in order for it to count as a valid

vote, then there would be no basis for counting ballots by hand to see if the

machines had failed to recognize valid votes. Republicans began to express this

view, arguing that the vote-counting process should be declared finished, once the

28

Electronic copy available at: [Link]

ballots had been run through the machines a second time, as required by Florida

law.

Gore’s attorneys, however, pointed to language in Florida law suggesting

that a ballot should be counted as long as it was possible for a manual

examination of the ballot to discern “the intent of the voter,” even if a machine

would be incapable of reading it. But this argument invited another question: how

could one be sure of the voter’s intent with respect to the wide variety of chads

that might not be completely dislodged? As it turned out, there were not only

hanging chads, but also dimpled or “pregnant” ones: chads that had been clearly

indented by the push of a stylus, and thus perhaps indicating the voter’s intent for

that candidate (especially if the voter was frail and the equipment that the voter

was using had become clogged with a pile of previously dislodged chad). These

dimpled chads, however, had not been punctured through at all, and thus perhaps

indicated hesitancy on the part of the voter, who decided in the end not to vote for

that candidate. Perhaps, then, there should be a requirement that a chad be

punctured in order to count as a valid vote. Yet even with respect to punctured

chads, not all ballots were the same. Some, indeed, where hanging by two

corners, or dangling by just one corner; others were barely punctured at all. If

Florida was to consider counting these ballots based on a manual examination of

them, should the state use a “two-corner rule” (requiring that the chad indeed be

hanging by two corners or fewer in order to be counted), as some other states did?

29

Electronic copy available at: [Link]

Or was it enough that there was any puncture to the chad at all, letting through a

little bit of light? Such questions might seem nitpicking, but they would prove

momentous when the dispute eventually reached the U.S. Supreme Court.

Meanwhile, Gore’s attorneys faced another practical problem in their

effort to seek a manual recount of these punch-card ballots. Florida law did not

provide for a statewide manual recount. Instead, using a “protest” procedure,

Gore was required to petition each county within the state for a manual recount of

that county’s ballots. Moreover, the “protest” rules seemed to give each county

little time to complete a manual recount: they were required to submit their results

to the Secretary of State within a week after Election Day. Alternatively, Gore

could let the result of the machine recount become certified and then file a judicial

“contest” of the result in state court. Gore, however, did not want Bush’s narrow

lead to become certified as an official victory. Therefore, he decided to go the

“protest” route and asked for a hand recount in four predominantly urban counties

where manual examination of the ballots would most likely yield the largest gain

for him.

The strategy backfired. Three of the four counties were unable to

complete a manual recount within the statutorily specified deadline. Gore’s

lawyers went to state court to ask for an extension of this deadline. Florida’s

Secretary of State Katherine Harris, a loyal Republican, announced that she would

certify final statewide results on November 18, after the deadline for the arrival of

30

Electronic copy available at: [Link]

overseas and military absentee ballots. (For political reasons, Gore decided to

abandon any possible legal challenges concerning the validity of military ballots

that might not comply with the requirements of state law.) Gore convinced the

Florida Supreme Court to extend the statutory deadline for completing the manual

recounts under the “protest” procedures until Sunday, November 26. But on

Wednesday, November 22, Miami-Dade County (Florida’s most populous)

decided to stop its manual recount, after Republicans held a rally claiming that the

county’s recount procedures were unfair. When the new November 26 deadline

arrived, Palm Beach County still had not finished its manual recount. Therefore,

as promised, Secretary of State Harris certified that Bush was the winner in

Florida by 537 votes.

At this point, Gore’s attorneys had no choice but to file the judicial contest

of the certification, which previously they had hoped to avoid. By this time,

however, the Safe-Harbor deadline under the Electoral Count Act, which would

require Congress to accept Florida’s resolution of the dispute if the state could

finish its proceedings on or before December 12, is only two weeks away. On

Monday, December 4, the state’s trial judge rejected Gore’s judicial contest, but

then on Friday, December 8, by a 4-3 vote, the Florida Supreme Court reinstated

the contest and remanded the lawsuit for a statewide recount of all “undervote”

ballots, where a machine may have missed a vote intended by the voter. In its

majority opinion, the Florida Supreme Court sidestepped the question of what

31

Electronic copy available at: [Link]

specific standard the recount should use to decide whether a ballot contains a

valid vote. According to the majority opinion, the general “intent of the voter”

standard in Florida law sufficed for the statewide recount to begin.

Bush’s attorneys immediately asked the U.S. Supreme Court to review the

Florida Supreme Court’s decision. The next day, Saturday December 9, by a 5-4

vote, the U.S. Supreme Court agreed to take the case and halted the recount

pending the outcome of its review. After hearing from the attorneys for both

sides on Monday, December 11, the following day—Tuesday, December 12, the

date of the Safe-Harbor deadline—the five-member majority of the U.S. Supreme

Court released an opinion finding, first, that the recount ordered by the Florida

Supreme Court violated the Equal Protection Clause of the Fourteenth

Amendment to the U.S. Constitution and, second, that any further proceedings

would be contrary to Florida’s self-professed desire to take advantage of the Safe-

Harbor Deadline. The majority of the U.S. Supreme Court therefore ordered that

no further recount take place.

The U.S. Supreme Court’s reasoning for its decision has been much

scrutinized since it was announced. Seven of the Court’s nine Justices agreed that

there was an Equal Protection problem, but only five signed on to the opinion

prohibiting any further recount. Justices David Souter and Stephen Breyer, who

accepted the majority’s premise of an Equal Protection violation, thought that the

Florida Supreme Court should be given an opportunity to decide whether

32

Electronic copy available at: [Link]

continuing with a constitutionally valid recount was more important than meeting

the Safe-Harbor deadline.

On the Equal Protection issue, the majority said that the central flaw was

that Florida law could have been more specific in providing when a less-than-

fully-dislodged chad is entitled to be counted as a valid vote. Because Florida law

remained unacceptably vague in providing no more than that ballots should be

counted according to the “intent of the voter,” different counties had adopted

different specific standards, and indeed the standard sometimes varied from time

to time, or place to place, within a single county. The majority gave the example

of Palm Beach County, which at first “precluded counting completed attached

chads, switched to a rule that considered a vote to be legal if any light could be

seen through a chad, changed back to [its earlier rule,] and then abandoned any

pretense of a per se rule, only to have a court order that the county consider

dimpled chads legal.” After reciting these facts, the high Court pronounced its

judgment: “This is not a process with sufficient guarantees of equal treatment.”14

The majority acknowledged that not every variation in voting practices

among localities within a single state would violate Equal Protection. Different

localities, for example, routinely use different types of voting technologies for the

casting and counting of ballots, and this difference has never been thought to

violate Equal Protection. Indeed, this kind of well-accepted variation among local

14
Bush v. Gore, 531 U.S. 108, 106-107 (2000).

33

Electronic copy available at: [Link]

voting practices is what caused Justices Steven and Ginsburg to dissent not just

from the majority’s decision to preclude any further recounting but also from the

majority’s predicate determination that an Equal Protection violation had occurred

in this case.

The majority opinion offered nothing further to explain the constitutional

distinction between permissible and impermissible variations in local voting

practices. Its opinion said only: “Our consideration is limited to the present

circumstances, for the problem of equal protection in election processes generally

presents many complexities.” This sentence has been the focus of much criticism.

Some have seen it as a confession from the Court that its decision was

unprincipled and intended only to end this particular presidential election with a

victory for Bush. More likely, it was just a signal from the Court that it was

treading very carefully in an area of law in which it did not feel sure-footed.

In any event, the murkiness of the majority opinion in Bush v. Gore leaves

the future of recount law very much uncertain. First, with respect to any disputed

presidential election in the future, there is the basic question of whether the U.S.

Supreme Court will involve itself again or instead will leave the dispute for

Congress to handle under the Electoral Count Act. Second, with respect to any

type of elective office, high or low, Bush v. Gore opens the door to a myriad of

possible new Equal Protection claims concerning the differential treatment of

similar ballots in the vote-counting process.

34

Electronic copy available at: [Link]

 As for the resolution of the 2000 presidential election itself, it is doubtful

that the U.S. Supreme Court’s decision actually made a difference in which

candidate would get to occupy the Oval Office. A study of all the state’s ballots

afterwards determined that Gore would have lost, even if the record had gone

forward under the most generous version of the “intent of the voter” standard (one

that would count any dimpled chad as a vote). Only if the manual recount had

also included “overvotes”—where a machine voided a ballot because it detected

votes for two or more candidates—would Gore have pulled ahead of Bush,

according to this media study. But Gore never asked for a manual recount of

overvotes, and the Florida Supreme Court’s order of a statewide recount extended

only to all undervotes and did not include overvotes.

Thus, although more Florida voters who cast ballots on November 7, 2000

may have wanted Gore to win—especially if one factors in the “butterfly ballot”

problem—the legal machinery of the state most likely would have ended with a

Bush victory even if the U.S. Supreme Court never halted the recount that was

underway.

Recounts since 2000—and in the future

After 2000, the nation eliminated punch-card voting machines and made

other reforms of the voting process, but did not alter the method for holding

35

Electronic copy available at: [Link]

recounts even in presidential elections. The Electoral Count Act of 1887 is still in

place unchanged (and essentially untested). The 2004 presidential election turned

on a single state, Ohio, but the popular vote in that state was not nearly close

enough—a margin of 118,601 votes—to fight about. The 2008 election was an

Electoral College landslide. Who knows how many times the nation will pick a

president before it suffers another dispute like the ones in 1876 and 2000? It

could be another century, or it could be much sooner (or perhaps even longer).

There have been two major non-presidential recounts since 2000. One

concerned the 2004 election for governor in the state of Washington. The other

involved the 2008 election for U.S. Senator from Minnesota. Both recounts

exposed significant weaknesses in each state’s vote-counting process. In

Washington, most egregiously, there were some 1600 unlawful ballots included in

the final certified count of the election, even though the margin of victory for the

winning candidate was little more than 100 votes, and thus the unlawful ballots—

mostly cast by former felons not entitled to participate under state law—may well

have determined the election’s outcome. Local election officials in Seattle had

also misplaced hundreds of voter registration forms and would have wrongfully

disenfranchised those voters if the forms had not been found during the recount.

In Minnesota, the major problem concerned absentee ballots: it turned out that

local officials had been treating differently absentee ballots affected by similar

clerical errors. Even so, Minnesota’s judiciary ruled that this differential

36

Electronic copy available at: [Link]

treatment did not violate the new Equal Protection principle of Bush v. Gore,

because the variation had been caused primarily by some local officials deviating

from the strict requirements of an unambiguous statute—not, as in Florida, local

officials operating on their own as a result of an insufficiently specific statute.15

(At least one federal court of appeals has interpreted Bush v. Gore differently—

saying, in the context of a dispute over a local judicial election, that variations in

vote-counting caused by mistakes may still amount to Equal Protection

violations16—and thus at some point the U.S. Supreme Court may need to sort out

the conflicting judicial interpretations of Bush v. Gore.)

Perhaps most troubling about the recounts in Washington and Minnesota

is the amount of time it took to resolve those two disputed elections. In neither

case did the dispute end until June of the following year (June 6, 2005 for the

Washington dispute; June 30, 2009 for Minnesota’s.) Thus, in neither case would

the state have been close to resolving the dispute by the day in December that the

Electoral College met that year to cast the official votes for president—much less

the by the Safe Harbor deadline, six days earlier. Thus, if either the Washington

or Minnesota recount had involved a presidential election, neither state would

have been able to finish it on time. Indeed, in both cases, if the recount had been

15
I have written on lessons to be learned from Minnesota’s 2008 election, as well as Washington’s
in 2004. See Edward B. Foley, How to Make Fair Faster, 10 ELECTION L. J. 187 (2011). For a
more extensive narrative of the 2008 election, see The Lake Wobegone Recount, 10 ELECTION L.
J. 129-164 (2011).
16
Hunter v. Hamilton County Board of Elections, 635 F.3d 219 (6th Cir. 2011).

37

Electronic copy available at: [Link]

cut short to satisfy the constitutionally mandated date for the Electoral College’s

meeting, the apparent winner of the election would have been the candidate who

ended up losing the recount eventually.

But there were good points to the Washington and Minnesota recounts.

For one thing, neither involved any threat of political violence. In both instances,

partisans supporting each candidate were willing to let the legal process run its

course and then abide by its results—a testament to our nation’s increased

commitment to the rule of law in the early part of the twenty-first century.

Moreover, to resolve the dispute over its 2008 election, Minnesota drew upon the

positive lessons learned from the successful recount of its 1962 gubernatorial

election: the state’s supreme court handpicked three judges to hear the dispute

over absentee ballots, and the way in which these three judges were selected

guaranteed that this special court would be perceived as impartial by both sides

and by the public at large. Although in this instance, the two candidates did not

choose the three judges, the state supreme court picked three judges from

different partisan backgrounds—including one demonstrating independence from

the two major parties—and thus the composition of the three-judge panel as a

whole was balanced and neutral.

Whenever another state next experiences a recount in a major statewide

election—and one shall surely occur, we just do not know exactly where and

when—that recount is likely to expose flaws in the state’s vote-counting process,

38

Electronic copy available at: [Link]

as did the recounts in Washington and Minnesota. In the field of election law, so

much attention is devoted to the rules governing campaign practices as well as

voter registration and the casting of ballots, or what we might call the “pre-

voting” rules, but comparatively little attention is given to the rules that govern

the counting of ballots, or what we can call the “post-voting” rules. The reason

for this discrepancy is that the pre-voting rules always matter to candidates and

their campaigns. In their efforts to win, one side or the other might gain a little

advantage by tweaking the pre-voting rules a little this way or that, and—you

never know ahead of time—a little tweak here or there might be just enough to

make the difference in a competitive race. But after the ballots are cast and the

initial count is in, it is usually unlikely that the content of the post-voting rules

will make a difference in the outcome of the election. Thus, the post-voting rules

do not receive the same rigorous scrutiny as the pre-voting rules and instead are

allowed to atrophy.

Moreover, there is little ideological incentive for political parties to fight

before Election Day over the vote-counting rules that will govern after Election

Day. The reason is that the parties do not know which vote-counting rules they

would prefer until, in each election, after the ballots are cast and they learn the

initial returns. At that point, the candidate who is behind will always favor rules

that make it easier to count more ballots, while the candidate who is ahead will

always favor just the opposite—a stricter enforcement of the rules that make it

39

Electronic copy available at: [Link]

more difficult to add previously uncounted ballots to the result. (This universal

truth applies regardless of whether a candidate is a Democrat or Republican, as a

comparison of the arguments in the 2000 and 2008 disputes demonstrates:

Democrats favored leniency in 2000, when they were behind, but advocated

strictness in 2008 once they took the lead in the Minnesota dispute; Republicans

urged strictness in 2000, but preferred leniency in 2008 once they fell behind.)

The lack of a partisan preference for particular vote-counting rules before

Election Day means, in principle, that it should be possible for Democrats and

Republicans in advance to agree upon a set of clear rules to govern the vote-

counting process. Everyone familiar with vote-counting disputes agrees that

having clear rules in advance to specify exactly what should happen in the event

of a new vote-counting controversy is a paramount value. Yet, precisely because

there is little incentive for politicians to focus on vote-counting rules, this

advantageous clarity usually does not exist.

Consequently, any state is likely to be in for a rough ride if it gets hit with

an exceptionally close race in a major statewide election. But from a national

perspective, no serious damage is done to democracy if every once in a while

some state somewhere around the country suffers a difficult recount. Of course,

one hopes the vote-counting dispute in any state never gets so bad as present the

threat of political violence, as occurred in some of the nineteenth-century

episodes. But the overall historical trend is to trust the state’s courts to settle such

40

Electronic copy available at: [Link]

vote-counting disputes, and that trend promises to reduce the risk of serious

political ill will, especially if states can adopt Minnesota’s approach of having

transparently fair tribunals adjudicate these disputes.

Finally, as for what might happen if and when the nation faces another

disputed presidential election, all bets are off. Neither Congress nor the states

have seriously examined the obvious gap—made even more evident by the

experience of Washington and Minnesota—between the time under the Electoral

Count Act that Congress allocates for the resolution of a disputed presidential

election under state law and the time that it actually takes a state to resolve a

disputed statewide election involving an important race. Nor have Congress and

the states taken steps to assure that the public would perceive the resolution of a

disputed presidential election to be fair and worthy of respect.

The nation was very fortunate that there was no political violence over the

outcome of either the 1876 or 2000 election. There was a serious risk of violence

over the Hayes-Tilden dispute. There were some calls for armed resistance, or

even another civil war, if Tilden was not declared the winner. Outgoing President

Grant even declared martial law on the eve of Hayes’s inauguration, in order to

prevent risk of insurrection. But in the end the transition of power was peaceful,

as it was on January 20, 2001. In addition to hoping that the next such

circumstance is also peaceful, the nation would be well served if Congress and the

states took steps to increase the likelihood that a vote-counting dispute in any

41

Electronic copy available at: [Link]

future presidential election would be resolved smoothly, predictably, and with

transparent fairness to both sides in the dispute.

In the meantime, perhaps the public—with the help of the media—can

become more comfortable with the notion that it may not know the decisive

winner of a presidential election immediately after the polls close on Election

Day. But the beginning of the primary process for the 2012 presidential does not

give one reason to be especially optimistic in this regard. For the Iowa caucus,

there was a rush to declare Mitt Romney the winner over Rick Santorum, by just

eight votes, on the same night that the caucuses were held. Even after the

canvassing of the caucus’s results was complete, and it was impossible to be

certain which candidate actually had prevailed because the official tallies from

some precincts remained missing, the media insisted on pronouncing Santorum

the decisive “winner” because he received more of the votes that could be tallied.

Ambiguity in the outcome appeared to be concept too difficult for the media to

grasp, even though ambiguity in the outcome was what the reality of the vote-

counting process had revealed.

The voters, one might surmise, are smarter than the media (at least the

televised media) give them credit for. It would improve the public’s

understanding of the vote-counting process—and thus, the public’s understanding

of its own democratic self-government—if the media helped the public achieve a

42

Electronic copy available at: [Link]

more sophisticated, and hence more accurate, appreciation for exactly when and

how official election results are determined.

43

Electronic copy available at: [Link]