Legal Instability

DeWitt and Schwartz say:

They assert that a state could withdraw from the compact “during a presidential election campaign” or “after the popular votes are counted but before the Electoral College meets.”

In fact, no state—whether operating under either a state winner-take-all statute or the National Popular Vote compact—may, after Election Day, change its method of appointing presidential electors and then appoint different presidential electors. The Constitution (Art. II, §1, Cl 4) gives Congress the power to “determine the Time of choosing the Electors” and Congress (3 USC §1) has designated one specific day (“the Tuesday after the first Monday in November”) as the day for appointing presidential electors. Federal law (3 USC §5) provides an additional backstop by treating a state’s appointment of its presidential electors as “conclusive” only if based on “laws enacted prior to” Election Day. Additionally, the notion of a state legislature enthroning the second-place candidate as president by changing the “rules of the game” after Election Day is not only politically preposterous, but contrary to law.

The compact actually provides more “durability” than the current system prior to Election Day because existing state winner-take-all laws are ordinary state statutes that a state has the power to repeal right up to Election Day.

In contrast, the compact does not allow its repeal to become effective during a six-month “blackout” period starting July 20 of a presidential election year and ending with the Inauguration on January 20. This six-month period includes the nominating conventions, the campaign, Election Day, the Electoral College meeting in mid-December, counting of votes by Congress in early January, and Inauguration Day.

DeWitt and Schwartz incorrectly assert that an interstate compact is not really binding and there is no sanction but the consciences of its founders. In fact, an interstate compact is not only a legally binding contract between states, but is one of the few ways by which a state legislature may bind future legislatures. The Constitution’s Impairments Clause (Art. I, §10, Cl 1) provides “No State shall … pass any … Law impairing the Obligation of Contracts.”

Once a state voluntarily enters into any interstate compact, a state cannot withdraw from the compact except in the manner permitted by the compact itself. As the US Supreme Court succinctly ruled, “A compact, is after all, a contract.” ^{Footnote 4}

The courts have never allowed any state to withdraw from any interstate compact without following the procedure for withdrawal specified by the compact itself. Federal and state courts have routinely and consistently enforced the US Supreme Court’s interpretation of the Impairments Clause as applied to compacts. ^{Footnote 5}

DeWitt and Schwartz incorrectly assert that an interstate compact is only enforceable by federal courts if it has received congressional consent. In the 1991 case of McComb v. Wambaugh (934 F. 2d 474), federal courts enforced a compact that did not require (and had not received) congressional consent (and that had a two-year delay on withdrawal) saying, “Having entered into a contract, a participant state may not unilaterally change its terms. A Compact also takes precedence over statutory law in member states.” Additionally, state courts have independent power to invalidate a state’s attempted withdrawal contrary to a compact’s withdrawal procedure.

Note that the compact’s six-month blackout period provides an additional independent constitutional impediment to DeWitt and Schwartz’s conjectured post-election repeal.

Uncooperative Electors

DeWitt and Schwartz’s statement that legislative support for the compact has been almost wholly Democratic describes a majority (but not all) of the early adopters of the compact. However, because the emotions associated with the 2000 election have now largely subsided, the most recent legislative floor votes were a bipartisan 40–16 vote in the Republican-controlled Arizona House, a bipartisan 28–18 vote in the Republican-controlled Oklahoma Senate, a bipartisan 57–4 vote in the Republican-controlled New York Senate, and a bipartisan 37–21 vote in the Democratic-controlled Oregon House. The most recent state legislative committee actions in 2016 were unanimous favorable votes by Republican-controlled committees in Missouri and Georgia. Polls in numerous states and nationally typically show about three-quarters of voters (including about two-thirds of Republicans) believe that the president should be the candidate who receives the most popular votes nationwide. ^{Footnote 6}

However, even if Republicans unanimously opposed the concept of the compact, no Republican presidential elector would, in the real world, vote against the Republican Party’s presidential nominee (who has won the national popular vote) because of a “principled preference for discrediting the compact.”

DeWitt and Schwartz incorrectly claim that a mere “handful of electors opposed to the compact could block its effect.” However, the compact is considerably less susceptible to disruption by faithless electors than the current system. The compact gives the national popular vote winner at least 270 electoral votes from the compacting states. In practice, the national popular vote winner would also receive a substantial cushion of electoral votes from non-compacting states that he or she won (on average, about half of the remaining electoral votes).

Finally, DeWitt and Schwartz incorrectly say there is no way to guarantee faithfulness of presidential electors. In fact, a state may simply follow the recommendation of the Uniform Law Commission and enact the Uniform Faithful Presidential Electors Act ^{Footnote 7} (which automatically cancels a faithless elector’s vote, removes that elector from office, and replaces the faithless elector with a loyal elector). Alternatively, a state could enact Pennsylvania’s existing law empowering each party’s presidential nominee to directly appoint his or her own presidential electors.

Manipulation of Vote Counts

DeWitt and Schwartz say:

In fact, the current state-by-state winner-take-all system greatly magnifies the payoff for vote manipulation in swing states—and hence the incentive to manipulate. A mere 537 popular votes in Florida in 2000 flipped enough electoral votes to decide the presidency in an election with a nationwide 537,179-vote margin. Astonishingly, while arguing that the current state-by-state system minimizes the incentive to manipulate votes, DeWitt and Schwartz cite a paper (Kosuke and King Reference Imai and King2004) in their own paper making a strong case that this 537-vote margin was the result of Florida’s illegal counting of 680 late absentee ballots.

DeWitt and Schwartz’s statement that the current system quarantines … manipulation of vote counts … within a small number of states offers little comfort, given the reality that this “small number of states” are the battleground states that actually decide who wins the presidency.

As former Senator Birch Bayh (D–Indiana) said,

DeWitt and Schwartz’s assurance that “partisan majorities … are not strong enough in closely divided states” is at variance with the fact that, as recently as 2012, one party controlled the law-making process in battleground states with over two-thirds of the general-election campaign events. ^{Footnote 9}

Narrowed Support

Critics of a National Popular Vote frequently offer hyperbolic predictions that presidential candidates would concentrate their campaigns on narrow segments of the population (particularly the big cities).

Even after acknowledging that the compact would make “every vote count, regardless of location,” DeWitt and Schwartz puzzlingly assert that it would “encourage candidates to appeal to narrower, less diverse interests” and encourage candidates to campaign in “fewer states.”

Given that every 2012 general-election campaign event was in just 12 states (and over two-thirds were in just four states), one wonders just how many “fewer states” DeWitt and Schwartz think would receive attention in a national popular vote for president.

The actual observed behavior of presidential candidates (advised by the nation’s most astute political strategists) inside battleground states indicates that they campaign broadly throughout the jurisdiction involved when the winner is the candidate receiving the most popular votes.

When real-world presidential candidates campaign in a state such as Ohio (which alone accounted for over a quarter of the nation’s general-election campaign events in 2012), they do not appeal to narrow interests; they do not concentrate on heavily populated urban areas; and it would be inconceivable that they would neglect three-quarters of the electorate. Instead, they campaign broadly throughout the state precisely because “every vote counts, regardless of location.”

Table 1 shows the four metropolitan statistical areas (MSAs) centered on Ohio’s biggest cities, the seven MSAs centered on the state’s medium-sized cities, and the 53 rural counties outside the MSAs. The third column of the table shows the number of campaign events (out of Ohio’s 73 events) that each part of the state would have received if the candidates conducted campaign events based strictly on the basis of the area’s share of the state’s population. The fourth column shows the actual number of campaign events for each part of the state. As can be seen, each part of Ohio received campaign events almost exactly in line with its share of the state’s population in 2012 (the tiny differences being the result only of rounding errors).

Similarly, campaigning inside Ohio is also distributed almost uniformly over the state’s 16 (equal-population) congressional districts. ^{Footnote 10}

Candidates campaigned similarly in the three other major battleground states (Florida, Virginia, and Iowa) which, along with Ohio, accounted for over two-thirds of all general-election campaign events. ^{Footnote 11}

Close Votes and Recounts

DeWittt and Schwartz claim the current system quarantines … recounts of close votes …within a small number of states.

In fact, the current state-by-state winner-take-all system does precisely the opposite. It has repeatedly created artificial crises that would not have occurred if the winner were simply the national popular vote winner.

The current system has generated five litigated state counts in our nation’s 57 presidential elections. It created an artificial crisis in 2000 because of the tiny state-level margin of 537 popular votes in Florida in an election with a nationwide popular-vote margin of 537,179. Similarly, it created an artificial crisis in 1876 because of small state-level margins (889, 922, and 4,807) in an election with a nationwide margin of 254,694. In Hawaii in 1960, there was litigation over a 115-vote state-level margin in an election with a nationwide margin of 118,574. Were it not for state winner-take-all laws, all five of these trivially small state-level margins would have been irrelevant almanac footnotes in elections with decisive six-digit nationwide margins. Far from being a quarantine, the current state-by-state system produces unnecessary infections.

Recounts are very rare occurrences in ordinary elections when the winner is the candidate receiving the most votes from a single pool of votes.

Among the 4,691 statewide general elections between 2000 and 2015, there were only 27 recounts—that is, a probability of 1-in-174. ^{Footnote 12}

This historically observed probability of statewide recounts can be used to estimate the probability of a national recount. One should expect a recount of a single-pool plurality-vote national election once in every 174 presidential elections—that is, once in every 696 years. In fact, a national recount is even less likely, because the probability of a recount diminishes with the size of the voting pool.

The reason why the current system gratuitously generates so many unnecessary disputes (5 litigated state counts in a mere 57 presidential elections) is that the nation’s 57 presidential elections were really 2,237 separate state-level elections—each a separate opportunity for a razor-thin state-level margin. Under the current state-by-state system, Russian Roulette is played 51 times every four years, whereas it would be played only once every four years under a national popular vote.

DeWitt and Schwartz also suggest that the appropriate trigger for a nationwide recount should be a 1% margin. “Popular votes with margins of less than one percent have occurred surprisingly often: in 1880, 1884, 1888, 1960, 1968, and 2000.” But a 1% nationwide margin corresponds to 1,300,000 votes in a present-day presidential election.

In reality, recounts shift very few votes. The average shift in the 27 statewide recounts conducted between 2000 and 2015 was only 282 votes. The distribution of these observed shifts has a standard deviation of 336. ^{Footnote 13}

Applying the Central Limit Theorem to the historically observed distribution of shifts resulting from statewide recounts, the average shift resulting from 51 statewide recounts (that is, a national recount) would be 14,382 votes (51 times 282). The standard deviation of the distribution of the shifts resulting from 51 statewide recounts is 2,400 (336 times the square root of 51). The mean (14,382) plus three times the standard deviation (three-sigma) is 21,582. The probability is 99.85% that the original apparent loser would not gain more than 21,582 votes in a national recount.

To put in perspective the inability of a shift of 21,582 to affect the outcome of a national election, recall that the closest nationwide margin (1960) in a presidential election in the past century was 118,574; the next three closest nationwide margins (1916, 1968, and 2000) were about a half million votes each; and all the remaining margins were in the millions. DeWitt and Schwartz’s claim that a 1,300,000-vote nationwide margin would warrant a national recount is therefore unrealistic.

Oblivious to the reality that the 537,179-vote nationwide margin in 2000 was far too large to warrant a recount if the compact had been in effect, DeWitt and Schwartz go on to paint an apocalyptic legal scenario of a national recount meandering for eight months after Inauguration Day in 2001. “Acting President Hastert would have presided over an agonizing nation-wide recount, a temporary administration, a recession, and maybe 9/11.” Unlike DeWitt and Schwartz, the courts and both parties to the litigation in 2000 were well aware of the constitutional imperative (Art. II, §1, Cl 4) that the electoral votes of all states must be finalized and cast on the “same [day] throughout the United States,” namely on December 18, 2000—regardless of whether desired or desirable recounts had been completed.

Both the courts and the 2000 litigants were also aware of the related statutory “safe harbor” deadline of December 12, 2000, for each state’s conclusive “final determination” of its canvas (as provided in the Electoral Count Act of 1887—now 3 USC §5).

Astonishingly, DeWitt and Schwartz go on to assert that, even if the 537,179-vote nationwide margin had survived a recount in 2000,

In other words, DeWitt and Schwartz claim that the public should not trust a 537,179-vote nationwide margin that survived a recount, while arguing that it should trust the un-recounted 537-vote state-level margin that actually decided the Presidency in 2000.

We agree with DeWitt and Schwartz that existing state recount laws are inadequate. However, these inadequacies are far more likely to cause trouble under the current dispute-prone state-by-state winner-take-all system than under a national popular vote. Because of the statutory schedule adopted after ratification of the 20^th Amendment in 1934, there are only five weeks between Election Day in November and the meeting of the Electoral College in mid-December. Because there cannot be a recount until there has been a count (that is, an initial official count), the on-the-ground reality is that no presidential recount has ever been completed on a timely basis prior to the meeting of the Electoral College. The only full presidential recount that was ever completed (Hawaii in 1960) was not completed until after the Electoral College met. Vice President Nixon, the loser of that recount (which reversed the original result), presided over the counting of the electoral votes in Congress and graciously allowed Hawaii to be credited to Kennedy, while simultaneously ruling that accepting the manifestly untimely recount did not set any precedent. Moreover, no presidential recount is ever likely to be completed on a timely basis because existing laws enable the leading candidate to simply “run out the clock” until the mid-December deadlines—exactly what happened in 2000. ^{Footnote 14}

Adoption of the National Popular Vote compact could well provide sufficient impetus for the states to upgrade their laws to guarantee a timely recount in a presidential election. Alternatively, Congress could (and should) exercise its existing power over the count in presidential elections (under the 12^th Amendment) and establish a national procedure for timely recounts along the lines of the draft law in Every Vote Equal (section 9.15.7 of Koza et al. Reference Koza, Fadem, Grueskin, Mandell, Richie and Zimmerman2013). Contrary to what DeWitt and Schwartz say, no overall federal “takeover” of elections (requiring a constitutional amendment) is required for Congress to exercise its existing power over the count. Because the current state-by-state winner-take-all system creates so many disputes, improving presidential recount laws is something that is more urgently needed by the dispute-prone current system than under a national popular vote (where a recount can be expected about once in 174 elections).

Implicit Restrictions Argument

Although the U.S. Constitution (Art. II, §1, Cl 1) provides, “Each State shall appoint, in such Manner as the Legislature thereof may direct, a Number of Electors….” DeWitt and Schwartz argue that the National Popular Vote compact might be unconstitutional because “The states themselves must still choose their electors … and they cannot devolve the choice on the other states.”

This attempt to read restrictions into Article II’s grant of power echoes the argument made by the losing attorney in McPherson v. Blacker—the preeminent case governing the power of a state to choose the method for appointing its presidential electors.

In its unanimous ruling in McPherson v. Blacker, the U.S. Supreme Court answered the arguments raised by the losing attorney as well as DeWitt and Schwartz.

In addition, the 10^th Amendment provides a rule for interpreting the Constitution when it comes to reading in limitations on the states’ exercise of their powers.

Small States Argument

DeWitt and Schwartz’s second constitutional argument is

In fact, state winner-take-all laws have already extinguished the political relevance of the small states in presidential elections. The closely divided battleground state of New Hampshire is the only one of the 13 smallest states (those with three or four electoral votes) that received any general-election campaign events in 2012. The irrelevance of the 12 other smallest states is highlighted if you notice that these states together have the same population (12 million) as battleground Ohio. These 12 smallest states together have 40 electoral votes—more than twice Ohio’s 18. Nonetheless, Ohio received 73 of the 253 campaign events, while these 12 smallest states received none. These 12 smallest states are totally ignored because they are all safe one-party states in presidential elections (six reliably Republican and six reliably Democratic). Under the state-by-state winner-take-all system, closeness—not size—matters in determining whether a state “wields disproportionate weight in the election of the President.”

Accordingly, in 1966, Delaware Attorney General David Buckson (R) led a group of 12 predominantly small states in suing New York (then a battleground state) in the U.S. Supreme Court in an unsuccessful effort to declare state winner-take-all statutes unconstitutional.

If making sure that “small states … wield disproportionate weight” were actually a criterion for declaring an election law unconstitutional, current state winner-take-all statutes would have been struck down by the courts long ago.

“Unconstitutional Vagueness in Violation of Due Process” Argument

DeWitt and Schwartz’s third argument as to why the compact might be unconstitutional is

If “generating more close elections and making it harder to tell who has won” were actually the criteria in American constitutional jurisprudence for declaring an election law unconstitutional, state winner-take-all laws would have long ago been struck down by the courts as unconstitutional (given their history of creating five litigated state counts in a mere 57 presidential elections—compared to the low 1-in-174 probability of a recount in an ordinary single-pool plurality-vote election).

Procedural Question as to Whether Congressional Consent Is Needed

Except for the relatively few compacts to which Congress has given advance consent, Congress does not consider interstate compacts until they have been enacted by the requisite combination of states.

Therefore, like virtually all interstate compacts, the text of the National Popular Vote compact is silent as to whether congressional consent is required in order for the compact to take effect. Thus, DeWitt and Schwartz’s claim that the National Popular Vote compact is ineligible to receive congressional consent because “the compact itself does not make it a condition of implementation” is simply incorrect.

After the National Popular Vote compact is enacted by states possessing a majority of the electoral votes, the compact provides that the governors of enacting states shall proclaim that the compact has taken effect. At that moment, opponents undoubtedly will initiate a lawsuit (likely by the attorney general of a non-compacting state under the Supreme Court’s original jurisdiction) contending that the gubernatorial proclamations were premature because the additional step of congressional consent is still required.

Although DeWitt and Schwartz suggest that the legal test for answering this procedural question depends on whether the compact is “minor” versus “sweeping” or whether it “affects … other states,” the actual legal test is provided in the U.S. Supreme Court’s ruling in U.S. Steel Corporation v. Multistate Tax Commission—the preeminent case governing whether a particular interstate compact requires congressional consent.

In McPherson v. Blacker, the U.S. Supreme Court answered the question of whether federal supremacy is threatened by a state’s choice of method of awarding its electoral votes.

Manifestly, when a state exercises one of its exclusive powers, it does not encroach upon or interfere with federal supremacy.

Moreover, even if there were some arguable second-order effect on federal power, the U.S. Supreme Court has specifically cautioned in U.S. Steel against

If the Supreme Court were to modify its existing precedents and require congressional approval, the compact could not take effect until subsequently approved by Congress. Such consideration by Congress would occur at a time when states representing a majority of the Electoral College had already enacted the compact.

DeWitt and Schwartz suggest that “standing” issues or timing problems would preclude the courts from deciding whether the compact was operative until after some future election in which some candidate lost because of the compact. In fact, the court system routinely settles election-law disputes on a pre-election basis (e.g., districting, ballot access, voting hours, voter identification).