There has been plenty of talk from the two leading Republican presidential contenders about how they will attract voters who didn’t embrace recent GOP nominees.

For Ted Cruz, his pool of new voters supposedly includes conservatives who didn’t bother to vote because they saw few differences between the parties. Donald Trump, on the other hand, promises to energize working-class voters who’ve been left behind.

For the sake of argument, let’s assume that both Cruz and Trump will attract some new voters, as well as some Democrats who simply cannot stomach Hillary Clinton, the all-but-certain Democratic nominee.

Unfortunately for Republicans, that doesn’t deal with the other side of the turnout equation: how GOP voters will behave.

Unable to support Trump and/or Cruz, some will stay home completely. Others will show up at the polls but leave the presidential race blank because they find Clinton and one or both of the current Republican front-runners unacceptable. Still others will defect to Clinton or another name on the ballot.

Obviously, defections are doubly damaging. Not only does the party experiencing the defection lose a vote, but the other party gains one.

Apart from the three-way races of 1992 and 1996 (when Republican presidential nominees drew just 73 percent and 81 percent of GOP voters), Republican defection rates have been remarkably consistent — and quite low — going back to 1980, according to exit poll data from Cornell University’s Roper Center for Public Opinion Research.

Only 6 percent of Republicans voted for Barack Obama in 2012, down from 9 percent who voted for him in 2008. Four years before that, 6 percent of Republicans defected to John Kerry, while 8 percent of GOP voters cast ballots for Al Gore in 2000, the same percentage who voted for Michael Dukakis in 1988. In 1984, 7 percent of self-identified Republicans defected to Walter Mondale, and 11 percent of Republicans voted for Jimmy Carter over Reagan in 1980.

The rates of Republicans defecting to the Democratic nominee over the past four presidential elections have ranged from 6 percent to 9 percent.

Democratic defections drop

Recently, Democratic defections have dropped dramatically from the 1980s, when they were very high.

In 1980, 27 percent of Democrats voted for Reagan, a single percentage point more than those who defected to him four years later. In 1988, 17 percent of self-identified Democrats voted for George H.W. Bush.

In 2000, Democratic defections to George W. Bush dropped to 11 percent, the same number that defected to him four years later. In 2008, 10 percent of Democrats backed John McCain, while only 7 percent of Democrats voted for Mitt Romney four years later.

The sky-high number of Democratic defections to the GOP nominee in the three elections held during the 1980s almost certainly reflects the realignment that was happening in the electorate.

Democratic voters who voted for Ronald Reagan (and other Republicans) but still considered themselves “Democrats” probably constituted the bulk of the defectors.

By the 2000 election, however, the ideological sorting of the two parties, aided by the strong wind of partisanship that swept over the country, had served to tamp down the numbers of partisan defectors. Still, Democrats are more likely to defect to the GOP nominee than are Republicans to defect to the Democratic nominee, though the differences have been relatively slight.

Of course, not all defections are equally damaging.

Republican voters who stay home or defect to Clinton in states like New York or Maryland will not affect the presidential contest, since those states’ electoral votes are never in play. But defections in swing states and even competitive ones could actually change electoral vote outcomes.

It’s impossible to know how many Republicans would vote for Clinton or sit out an election in which their party nominates either Trump or Cruz. There is also no way to know exactly how many new voters Trump or Cruz would bring to the GOP.  (See Karl Rove’s take on Cruz’s turnout argument in a 2015 Wall Street Journal column, as well as a different view from Philip Bump in The Washington Post.)

Numbers don’t bode well

Still, there are more than a few reasons to believe that the net outcome would not be good for the Republican Party.
The March 17-20 New York Times/CBS News poll asked each party’s primary voters whether they could support all of the candidates in the general election. While 10 percent of Democratic primary voters said they could not support Clinton if nominated, 17 percent of Republican primary voters said they couldn’t support Trump and 19 percent said they couldn’t support Cruz if nominated.

In Wisconsin’s GOP exit poll, a stunning 38 percent of Republican primary voters said that they would be “scared” if Trump is elected president, while another 20 percent said that they would be “concerned.”  In contrast, only 12 percent said they’d be “scared” if Cruz is elected and 25 percent said they’d be “concerned.”

On the Democratic side, only 7 percent of Wisconsin Democratic primary voters said they’d be scared if Clinton was the nominee, and 23 percent said they would be concerned.

In the Ohio Republican exit poll, 28 percent of respondents said that they would not vote for Cruz if he is the Republican nominee, while 31 percent said that they would not vote for Trump.

These numbers probably are inflated by the combative nature of the nominating process — in May 2008, many Clinton voters said they wouldn’t vote for Barack Obama — but even if half of those respondents who said they wouldn’t vote for Cruz or Trump bolted from their party’s nominee as promised, it would surely be fatal for the Republican nominee.

Finally, Trump’s very weak showings against Clinton in hypothetical ballot tests over the past six weeks — trailing by anywhere from 6 points to as many as 13 points — suggest significant numbers of GOP voters won’t support Trump in a general election. (Mitt Romney lost his bid in 2012 by 3.86 points, close to half of John McCain’s 7.28 point margin.)

Cruz doesn’t fare nearly as poorly as Trump in general election polling, but he would lose some Republican voters because of his uncompromising approach and his conservative social issue agenda.

Because of changes in the electorate and the controversial nature of the Democratic nominee, Republican defections from Trump probably would not approach 1964 or 1972 landslide levels. But defections certainly could double from the recent norm — jumping to 12 percent to 18 percent — levels that would make for a very difficult Republican year.