Given the tumultuous primary elections and caucuses dominating today’s headlines, many analysts and citizens are marveling at the byzantine process by which America picks its presidents. As we reflect on these matters, we should also take some time to think fundamentally about how we select vice presidents. (Indeed, the time is particularly ripe to focus on vice presidential selection, since presidential aspirants may try to increase their chances of succeeding at an “open” convention this summer—by which I mean a convention in which no presidential nominee has a majority of delegates firmly committed going in—by identifying in advance their preferred running mates, the way Ronald Reagan did in 1976, the last time the Republicans had an open convention.)
The very term “running mate” should, upon reflection, remind us that vice presidential elections are very odd in American democracy, insofar as voters are prevented from registering their preference for a president of one party and a vice president of another.
Consider the elections of 2008 and 1988. In those years, the Republican party picked individuals (Sarah Palin and Dan Quayle, respectively) who were considered by many Americans as underqualified to become president, and perhaps likely to lose in a head-to-head race against the veep candidate of the other party. (I should emphasize that some of the criticism of both of these individuals was unfair; I am not suggesting that the attitude of Americans was fully justified, only that it existed.) All of this raises an important question: Why were voters’ hands tied in this respect? Why, in these elections, were voters not permitted to select a John McCain/Joe Biden or George Bush/Lloyd Bentsen White House? Consider, by comparison, that voters are allowed to and often do split their votes between the Democratic and Republican parties in other ways. Sometimes it makes sense to vote for a president of one party and a House member or Senator of the other. Not infrequently in the modern era, voters split their tickets between federal officials of one party and state officials of the other. Indeed, in a number of states, voters are free to vote for a governor of one party and a lieutenant governor (who becomes governor if and when the governor is unable to serve) of another.
What explains the unusual limitation on voter freedom embodied in vice presidential selection? Importantly, nothing in the federal Constitution or federal election law requires that voters in each state choose between internally unified Republican and Democratic tickets. And having a president of one party and a vice president of the other is not lacking in historical precedent; in 1796, Thomas Jefferson was elected to serve as John Adams’s veep, even though the two men—bitter rivals—represented and headed opposing parties. Some early Americans thought that having a “split” executive promoted checks and balances, and was something to be embraced, not lamented. Interestingly, there was no great movement to amend the Constitution in the wake of the 1796 election to prevent a recurrence of a divided White House.
As Akhil Amar and I explained in a Virginia Law Review article, the adoption of the Twelfth Amendment eight years hence (in 1804) did make it easier for political parties (a formidable reality by the mid-1790s, even though their existence at the constitutional founding was neither guaranteed nor prohibited) to elect an aligned president and vice president, but nothing in the Amendment requires states to tie voters’ hands the way all states currently do. Instead, over the last two centuries, with the increased strength of the national political party apparatuses, each state has simply chosen to structure its presidential election ballot so as to require voters to pick between opposing slates—rather than allowing voters to choose separately for president and vice presidential candidates. (Part of the explanation here is that ticket splitting among voters more generally was not common until a few generations ago, and in an earlier era when voters rarely voted for candidates of more than one party, the question of unbinding the electorate so that it could pick a president and veep of different parties may not have been worth thinking about.)
Some decisions by states as to how to structure the mechanics of federal elections can be explained easily by the desire each state has to maximize its own clout. For example, (almost) every state’s decision to award its electoral college votes on a “winner take all,” rather than proportional, basis is explicable by reference to each state’s desire to reward the presidential candidate who takes that state’s concerns and issues most seriously. But the decision to bind voters to executive party tickets cannot so easily be explained by rational state selfishness. Any or all states could, if they wanted to, begin untying the voters’ choices in this regard, without causing any obvious harm to the states that so decide.
Some might argue that a state should bind its voters’ hands here because many voters in the state would prefer a unified presidential ticket of either party, rather than a split ticket. But the same might be true of ticket-splitting between the presidency and Congress; a small number of voters can effectively send a split message from a state even if a majority of voters would prefer a unified message in either direction. And, of course, a state could, if it wanted, structure its ballot so that voters’ second-choice preferences are given effect. This could be done by having a ballot that first asks for presidential preferences, and then asks for vice presidential preferences depending on who wins the presidency so that voters who have an “I’d prefer a straight Democratic presidential ticket but I would rather have a unified Republican ticket than a president of one party and a vice president of the other” mindset would have their preferences registered.
Of course, even if voters were given the option of White House ticket-splitting, they might not invoke that option lightly; there might be strong policy reasons for voters to generally prefer having a president and vice president of the same political party. For instance, a general preference in favor of policy consistency within a four-year period in the event of a presidential vacancy, or a desire not to encourage the incentive for political assassination, might counsel in favor of a unified president/vice president team. But those policy reasons could be overcome in an unusual year with unusual vice presidential candidates. (I should also note here that the practice of ticket “balancing,” wherein each presidential candidate picks a running mate to placate a different wing of the party, already creates some potential for policy inconsistency over a four-year term, as happened during Andrew Johnson’s succession in the wake of Abraham Lincoln’s death, and also some “reason” for a crazy person to assassinate a president, as in the case of James Garfield’s assassin, who said he committed the crime so that Chester Arthur, who came from a different strand of the party, would exercise power.) Generally speaking, with respect to policy reasons in favor of a unified White House, our traditional preference in American democracy is to let voters—rather than government—weigh competing policies and personalities.
Importantly, in the end, if voters had the power to split their executive votes, they might never need to exercise it. If states didn’t bind voters to party tickets, then Senator McCain could, as a practical matter, never have chosen the untested Governor Palin as his running mate; he would have anticipated that the Palin choice, rather than enhancing his ticket, would trigger a possible McCain/Biden win instead, teeing up Biden, a Democrat, to be in the line of presidential succession. Instead, Senator McCain would likely have chosen someone who was a stronger, nationally credible, candidate.
In addition to improving the quality of vice presidential candidates, untying voters’ hands might also reduce the incentive to “balance” the ticket across the various wings of the party, because a veep candidate who appeals to some of the party but not its mainstream might risk losing against a candidate from the other party with broader appeal. Ironically, then, unbundling presidential and vice presidential elections could end up generating vice presidents who are not just better qualified and more mainstream, but also more unified with the presidents alongside whom they serve.
Finally, and most importantly, separating presidential and vice presidential voting could give the elected vice presidents more credibility and legitimacy should they ever ascend to exercise the powers of the presidency (either temporarily or for the balance of an elected president’s term), or should they ever run for the presidency down the road. Right now, veeps have very little independent electoral credibility, which may hold them back in the kinds of tasks they perform.
Changing the way presidential ballots are structured is obviously a big deal, and the question of electing presidents and vice presidents separately is not one on which I’ve come to rest. But I do think (and have thought for some time) that it is a question worth examining carefully, as each passing election gives us information about how we might improve American democracy.