vision 2020

Does Labor Day Mark the Beginning of the Real Presidential Race?

John F. Kennedy in Detroit’s Cadillac Square on Labor Day 1960. Photo: Preston Stroup/AP/Shutterstock

For presidential candidates who end August trailing their opponents, there’s solace in the hoary maxim that campaigns really begin on Labor Day. Once upon a time, that was to some extent true. Candidates had big events on Labor Day, with Democrats traditionally holding a rally in Detroit’s Cadillac Square, as Decision Desk HQ’s Nick Field recently recalled:

In 1948, Harry Truman began his legendary comeback with a Labor Day kick-off in Cadillac Square, Detroit. After that, it became a tradition for every Democratic presidential nominee to begin their crusade there with a Labor Day stemwinder. 

This custom was ended by Jimmy Carter, who chose to start the 1976 fall campaign in Warm Springs, Georgia (a location meaningful to both himself and FDR) over the home state of his opponent Gerald Ford. 

And it wasn’t just Democrats who waited until Labor Day to get serious, either, as the Washington Post noted in a look back at the 1964 campaign:

[T]here was a time in American politics when general election presidential campaigns formally began AFTER Labor Day. [Barry] Goldwater received the Republican nomination for president in July, took the month of August off and then started his campaign on Sept. 3.

In more recent years the growing power of television and the spread of mass summer vacations reinforced the idea that presidential campaigns at least paused between the national political conventions and Labor Day. The quadrennial Olympics were one distraction, and it was generally assumed that voters were focused on other summertime leisure activities instead of the blood sport of politics. It certainly helped that Congress was almost invariably in recess in August, and that big-foot journalists were off to the Hamptons and Martha’s Vineyard where at most they might file airy columns on the mood of the nation.

The idea that “nothing happens in August” took a big hit in 2004, when a shady conservative group calling itself Swift Boat Veterans for Truth sponsored a series of four TV ads attacking the war record of Democratic presidential nominee John Kerry. The ads were widely thought to have changed the momentum of the contest and contributed to George W. Bush’s narrow November win, particularly after Kerry’s campaign was slow to respond to the assault.

Four years later Democrats held their national convention in late August, and then Republicans actually began theirs on Labor Day itself. By then both campaigns were in full operational mode.

But the belief remained strong in many circles that general election developments before Labor Day were strictly preliminary because voters had not yet “tuned in.”

There’s not actually much empirical evidence for the proposition that presidential elections are really shaped by post-Labor Day developments. A Gallup analysis of contests from 1936 through 2004 showed that the Gallup Poll leader on Labor Day ultimately won the popular vote in all but two cases: 1948, when there was little or no polling to document the Truman comeback, and in 1980, when Ronald Reagan blew by Jimmy Carter mostly because the voters mulling third-party candidate John Anderson predictably returned to the major-party fold as Election Day approached.

Since 1980 late shifts in the presidential race have been limited, with the exception of 1996, when Bill Clinton’s vast lead on Labor Day declined to a mere nine-point wipeout.

On Labor Day 2008, Barack Obama led John McCain in the RealClearPolitics polling average by 6.4 percent; on Election Day he won by 7.3 percent. On Labor Day 2012 Obama led Mitt Romney by a tiny 0.1 percent in the RCP averages, and led only by 0.7 percent in the final averages (though he actually won by 3.2 percent). And on Labor Day 2016 Hillary Clinton’s 3.3 percent lead was almost identical to her 3.2 percent lead in the final RCP averages; she actually won the popular vote by 2.1 percent.

One type of presidential campaign event of fairly recent vintage that does almost always happen after Labor Day is the candidate debate, or debates. These spectacles typically draw large audiences, and there is always the chance that a candidate gaffe (e.g., Gerald Ford’s insistence in 1976 that Poland was “free”) or some brilliant jab (Ronald Reagan’s “are you better off than you were four years ago” conclusion in 1980) will change the course of history. But political science research suggests that their effects are usually marginal.

And so the idea that presidential general elections begin on Labor Day denies all the evidence we have that nearly all of them were actually decided earlier, at least in terms of the national popular vote winner. That’s become even truer as partisan polarization has reduced the number of swing voters and hence the volatility of presidential elections, and early voting has reduced the time available to persuade and mobilize voters before they cast ballots. Sure, very close elections are subject to very small changes down the stretch, and as we learned in 2000 and 2016, the gap between popular and electoral vote totals means that popular vote leads may not be as invincible as they look. But you really don’t want to be a candidate facing a steep uphill climb two months from Election Day. That may be one reason Donald Trump’s campaign seems to be thinking about pulling a fast one with a premature claim of victory on Election Night.

Does Labor Day Mark the Beginning of the Presidential Race?