The way we pick VPs is terrible. The alternatives are even worse.

With Donald Trump and Hillary Clinton expected to announce their vice presidential nominees in the next week or two, the media is frantically speculating about whether their picks will help — or hurt — their electoral prospects.

Will Trump try to shore up his support among evangelical conservatives? Or will he pick an "attack dog" who will tear into Clinton? Will Clinton want a progressive firebrand to win over Bernie Sanders die-hards? Or will she attempt to cut into Trump’s big lead among white men?

What everyone takes for granted is that both Clinton and Trump aren't thinking all that much about who would be a good successor should they die in office. Instead, they’re at least partially thinking about who will best help them win this fall.

Vox's Matt Yglesias has a good critique of this line of thinking, and makes the case that the VP pick should be done with governing in mind, rather than short-term electoral politics. All this got me thinking, though — why should we force presidential nominees to choose their successors in the heat of an election?

Tom Williams/CQ Roll Call/Getty

One obvious argument for keeping the VPs on the ballot is that it gives voters a possible veto over particularly objectionable choices. But given that voters don't seem to factor the VP nominees into their choices very much, why shouldn’t the winner of the presidency make his or her pick after — rather than during — the heat of the race? Do we really want to risk letting the country fall into unsteady hands simply because Iowa looked like a battleground state one year?

On Tuesday, I talked to six political scientists who have studied the vice presidency in search of an answer for this question. It turns out there's actually a pretty good defense for keeping things as they are — just not the one I had expected.

The case for changing how America picks its vice presidents

The Constitution has required that the vice president be on the ballot, in one form or another, since our nation’s founding. But I was relieved to learn that I'm not the only one to think the way we pick VPs in the heat of an election cycle seems somewhat nutty.

Back in the early 1970s, Michigan Sen. Bob Griffin proposed an amendment to the Constitution that would have ended the direct election of vice presidents and instead let presidents appoint their VPs, subject to congressional approval, after being sworn in.

Of course, the idea didn’t go anywhere. But it wasn’t without its supporters — especially after Richard Nixon’s veep, Spiro Agnew, had to resign over accusations of extortion and tax fraud.

At the time, reporters like Tom Wicker of the New York Times made a case similar to mine, according to John D. Feerick’s The Twenty-fifth Amendment: Its Complete History and Applications. Wicker wrote:

The fact is that Presidents dictate their choices for Vice President with ruthless disregard for any[thing] but political qualifications, and political conventions approve them with no 'inquisition' at all. The further fact is that no Vice President is really elected by the people or their electors; he is an appendage to their Presidential choice.

Wicker’s point is twofold: 1) that VPs are often chosen out of "ruthless" political calculation alone, and 2) that voters don’t really get to vote on whether they support the VP.

Some of the contemporary political science research backs Wicker up. "There’s little evidence that voters make any consideration based on the vice president," says Joseph E. Uscinski, a political scientist at the University of Miami and the author of several studies on the vice presidency.

You’d think that nominees wouldn’t then let electoral factors seep into their choices. "But they do, and sometimes that gives us vice presidents who don’t do what the presidents, or the presidents' voters, wanted," Uscinski says.

History shows us that our VP system can go very wrong

You don’t need to go back to the 1970s to see how short-term electoral considerations can warp presidential candidates’ VP picks — with really worrying implications.

The most obvious recent example is Sarah Palin in 2008, when Sen. John McCain — hoping for a Hail Mary comeback in the polls — picked the obviously unqualified Alaska governor in the hopes of firing up his party’s base.

McCain and Palin lost, of course. But McCain certainly theoretically could have won, and he could have died while in office. So there’s a pretty good argument that we risked a potentially disastrous presidency in part because we make presidents choose their successors when an election is at stake.

There’s enough historical evidence of this happening for us to know that it isn’t just an unlikely possibility. In 1864, Abraham Lincoln chose southern white Unionist — and, it turned out, white supremacist — Andrew Johnson for his VP, in a bid for geographic diversity. After Lincoln was assassinated, Johnson advanced a disastrous set of Reconstruction policies totally at odds with what the "great emancipator" and his Republican base wanted. (Plus, Lincoln lost the Southern state of Kentucky that year anyway.)

Discrepancies between the POTUS and the VP has also made leadership crises more difficult to resolve, and may even create an incentive for political assassins, Uscinski wrote in one recent paper. In 1881, President James Garfield was shot twice in the back but stubbornly clung to life for 80 days while completely incapacitated. Some lawmakers wanted to resolve the crisis by making VP Chester Arthur president but feared the political implications of doing so, because Arthur and Garfield came from different factions and had completely contradictory policies on civil service reform.

"Presidential candidates have often opted for running mates who are politically different from the head of the ticket ... to increase the chances of winning the general election," Uscinski writes. "But when presidents and vice presidents differ significantly, the country could inherit espousing policies the country does not democratically support."

Uscinski’s argument here isn’t just that an "unqualified" VP could become president despite not enjoying the support of much of the country. It’s that we explicitly look to vice presidents to complement the ideological profiles of the nominees, thereby intentionally inserting confusion into our government that could, potentially, be avoided under a different system.

But what’s the alternative — giving Congress a veto over who the VP is?

So if the vice presidency is a powder keg of risk and undemocratic outcomes, then why shouldn’t presidents just appoint VPs like Cabinet members? (I know Vox stories are unlikely to spur new constitutional amendments, but bear with me.) After all, you never hear of a president tapping an unqualified secretary of state merely to cement a voting bloc these days — right?

"I agree that it’s a messy system and that there are a number of anomalies," says Joel Goldstein, a law professor at St. Louis University and the author of a new book about the modern vice presidency. "But my bottom line is that the system we have makes the most sense when all things are considered."

Goldstein has lots of compelling defenses of our current system, but they boil down to one essential point: Requiring VPs to be confirmed by a gridlocked Congress (as is the norm for high-level presidential appointees) could be nightmarish.

Right now, a president basically gets to pick anyone he wants for his running mate (if his convention’s delegates approve the choice). If he wins a landslide at the ballot box, the other party doesn’t get a veto of the VP. That would probably have to change if the US tried to convert the vice presidency into a position chosen after the election.

And just imagine a recently elected President Clinton trying to get her VP nominee through a Republican Senate, Goldstein says. The difference between who she thought would be good for the job and who she could get approved would, presumably, be even greater than it is under our current arrangement.

"Giving Congress a veto seems less democratic than giving the public a veto," Goldstein says.

He adds that vice presidents have become increasingly powerful over the past several decades. Subjecting that office to our current partisan warfare just doesn’t seem like a promising way to improve the VP’s accountability or effectiveness.

"The vice president has become an office that provides high-level help to the president as a senior political adviser and high-level troubleshooter," Goldstein says. "And that’s at a time when most of the rest of the government is appearing increasingly dysfunctional. I don’t think we should mess around with it."

The case that our current VP system actually gives voters a pretty good check

The upshot of the political science research is clear: Most of the media obsession over the VP selection is dramatically overstated.

In one study recently highlighted in Politico, two political scientists found that vice presidents don’t even have a statistical impact on increasing votes for the ticket in their home states.

"Demographic groups don’t fall in line because of a VP selection, and neither do states," says Kyle Kopko, a political scientist at Elizabethtown College, who conducted the research with Mount Vernon Nazarene University political scientist Chris Devine. "Voters think of partisan preferences and what the presidential candidates are saying."

But the political scientists also argued that the VP choice can make a difference — even if it tends not to. The experts can’t often find examples of VP picks really siphoning votes away from the nominee. But that may because our current system is effectively deterring presidential nominees from picking running mates that really hurt their chances.

Another reason for optimism: There’s growing evidence that presidential nominees understand the importance of picking a good governing partner over an illusory electoral gain, says Matt Grossmann, a political scientist at Michigan State.

"I think presidents are recognizing what the VP means," adds Matt Dickinson, a political scientist at Middlebury College. "The most recent presidential candidates seem to be made from a need for a governing partner as much as for any electoral gain."

The idea of Sarah Palin becoming president because of some quirk in the system seems to argue against this. But maybe she’s also a sign that voters really do punish — and therefore consider — terrible VP picks. (Indeed, one study has found that picking Palin cost McCain about 2 million votes.)

"Yes, there are some exceptions, but the fact that the presidential candidate has to pick someone before an election tends to force them to pick pretty good people," Goldstein says. "You’d really lose something by not having the vice president as part of the campaign."