Every four years, we seem to hear that we are facing the “most important election ever,” but this year that hyperbole has reached new heights. At the Republican National Convention, former New York City Mayor Rudolph Giuliani said starkly: “There’s no next election. This is it.” A conservative essayist in the Claremont Review of Books went even further, calling 2016 the “Flight 93 election” — the final moments in which to save a hijacked republic. And Roger Angell, the longtime New Yorker editor and Hillary Clinton backer, wrote recently that the coming election will be the most important of the 19 presidential elections in which he will have voted.
It’s easy to see why so many voters feel so strongly about the importance of the coming election, but as an empirical matter, statements like Giuliani’s are incorrect. That’s because, in comparison to many other developed democracies, the United States actually has frequent federal elections. In Canada and Britain, a single party can govern for up to four or five years before voters get to weigh in. In the U.S., by contrast, no party can maintain unified control of the federal government for more than two years before facing the voters. After 2016, we’ll have midterm elections in 2018, less than 22 months after the next president is inaugurated. I certainly won’t call those midterms the “most important ever,” but they will have a particular importance: Control of the U.S. Senate, and possibly the House, could hang in the balance. For whichever party that loses the 2016 presidential race, 2018 is a big-time consolation prize.
Since the 1930s, one of the most dependable regularities in American politics1 has been midterm loss, a swing against the party of the incumbent president. Whether due to a reaction to the sitting president’s agenda or to voters seeking a counterweight to the president, the party not holding the presidency has made gains in the House in the midterm elections in every election but two since 1934.2
The chart below shows every U.S. House election since 1952 according to the change in each party’s share of the vote from the election two years prior.3
The results are striking. What first jumps off the page is that the biggest “thumpings” and “shellackings” (to quote former President George W. Bush and President Obama, respectively) are almost always in midterm years — and almost always directed against the president’s party. The far-left column shows the six elections in which the GOP’s vote share dropped more than 4 percentage points. Republicans held the presidency in four of those years: 1958, 1974, 1982 and 2006. In 2006, the Republicans lost the House; in the other three elections, they didn’t hold it in the first place. Likewise, the two far-right columns show three years in which the Democrats saw their House vote share drop 6 to 10 percentage points: 1966, 1994 and 2010. They lost control of the House in two of those three elections. In fact, the last four times the House of Representatives changed hands were all midterm years: 1954, 1994, 2006 and 2010.
By contrast, look at the two middle bars representing swings of up to 2 percentage points. There, we find most of the presidential election years. Volatility in House elections is markedly higher in midterms. When studying politics, results are rarely this clear-cut. If history is any guide, the House GOP’s majority would face more risk from a President Donald Trump in 2018 than from Trump’s campaign in 2016.
Or imagine that Clinton prevails in November. If so, the most likely outcome is a continuation of the recent pattern of resounding GOP victories in midterm years. In 2018, Democrats will defend Senate seats they won in 2012 in several red states, including Indiana, Missouri, Montana, North Dakota and West Virginia. West Virginia’s Joe Manchin famously shot a copy of the cap and trade bill in a 2010 campaign advertisement — but even excellent marksmanship might not be enough in a third consecutive midterm wave for the GOP.
In addition, some of the most important aftershocks of the 2016 election are likely to be felt not in Washington but in state capitals across the country. In 2018, 36 states will choose governors. As I’ve pointed out before, our elections for governor increasingly track national trends. Governors are typically powerful officials in their own right, with substantial control over state budgets and policy. But even for those who care about power only at the federal level, there is good reason to care about the 2018 governors’ races: In many states with multiple House districts, those governors will have veto power over their states’ redistricting processes after the 2020 census. Over the course of the Obama presidency, anti-Obama voting in non-presidential years is a major reason why the Democrats have lost a net of 11 governors’ seats.
Likewise, 2016 has critical implications for state legislative elections. Political scientist Steven Rogers has shown that presidential approval is a powerful predictor of voting in state legislative races. Since Obama became president, the same dynamics have cost the Democrats approximately 818 seats in state legislatures, and they have lost control of 29 net chambers in state legislatures. Sure, holding the presidency allows a party to pursue its agenda at the federal level. But in recent decades, that pursuit has come at a remarkable down-ballot cost for Democrats and Republicans alike.
To explain why the electorate has alternated between leaning Democratic in recent presidential years and Republican during midterms, Obama argued that Democratic-leaning constituents are less likely to vote in midterm years. There’s some truth to that. But it’s not the main force behind the recent swings, as FiveThirtyEight’s Harry Enten has shown. Think about the math: Each voter who sits out a midterm costs his party one vote, while each voter who switches parties adds a vote to the new party while taking one away from the old party. The more powerful engine for change is that voters are changing their minds — and for decades, they’ve leaned against the party holding the White House.
To date, 2016 has defied many analysts’ expectations, and 2018 may as well, especially if either of the major parties divides in a consequential way. But with both major-party candidates being viewed unfavorably by record numbers of Americans, there is every reason to expect that 2018 will follow the trend of 2006, 2010 and 2014, with a decisive shift against the president’s party — whichever party that is.
University of Pennsylvania students Max Kaufman and Thomas Munson provided research for this article.