With another set of astonishing state poll numbers out this week, the ultimate outcome in the presidential race seems pretty well determined. Sure, history says that in a normal election cycle Republican nominee Donald Trump should be expected to gain back some of the percentage points lost after his disappointing convention and the Democrats’ successful one, but even standard tightening will likely leave him well short of victory (not to mention that normal events should not be presumed when it comes to Trump). For goodness sakes just look at the latest NBC/Marist poll numbers from four alleged battlegrounds, now swing states in lazy media usage only. Yes, in Florida Clinton is “only” ahead by 5 points, within striking distance should Trump resuscitate his campaign. But in states right at the tipping point of national averages, Clinton leads Trump by 9% in North Carolina, 13% in Virginia, and a remarkable 14% in Colorado. These are hardly outliers, either, though they are on the high side of current polling. The Real Clear Politics average for these states now has Clinton ahead by 8% in Virginia and 11% in Colorado. For comparison, in 2012, Barack Obama won Colorado by 5%, Virginia by 4%, and lost North Carolina by 2% as he was claiming a national popular vote mandate of nearly 4%. If Clinton holds on in Colorado and Virginia, along with defending two other traditional Democratic-leaning battlegrounds in which she has double digit leads (Pennsylvania and New Hampshire), her path to 270 electoral votes is without obstacle.
Assuming one cares about life after the election—as opposed to participating in presidential election hype as a cathartic identity exercise or a tactic for combatting summer dog-days boredom—then the real drama of the next three months is how these numbers will affect down-ballot congressional races. Traditionally, there is a correlation between a party’s success at the top of a ticket and its ability to make gains in Congress. Presumably this phenomenon is based on the increased attention and importance of the presidential race. If a presidential candidate can draw its voters to the polls in significant numbers, those voters will be more likely to vote for the president’s party in Congress. In 2008, Barack Obama’s 7% victory brought in eight new Senate seats and twenty-one new Democratic members of the House. Similarly, in 1980, a resounding win for Republican Ronald Reagan over incumbent Jimmy Carter led to Republican gains of twelve Senate seats.
At the same time, the popular mythology of the “coattail” effect often fails to materialize. Bill Clinton won significant victories in 1992 and 1996 with no change in Senate composition in the former year and losing two seats in the latter one. Richard Nixon, in winning reelection by one of the largest margins in the nation’s history in 1972, took only an additional thirteen House seats and lost two in the Senate, in what became known as the “lonely landslide.” The Milwaukee Journal Sentinel wrote the morning after that campaign: “President Nixon’s failure to pull in substantially more Republican members of Congress on the strength of his historic landslide Tuesday is the phenomenon of the 1972 presidential election.” Similar, though less pronounced, lonely landslides occurred in Lyndon Johnson’s win in 1964 and Reagan’s 1984 reelection.
So, significant presidential victories sometimes yield big gains in Congress and sometimes do not. Now, there is a burgeoning bandwagon of political pundits arguing that Trump’s unique candidacy won’t be a singular drag down-ballot. There is an emerging consensus that Republicans may be able to limit their losses even if The Donald takes a drubbing.
Examining current polling data, elections analyst Charlie Cook recently sounded strong notes of skepticism that a down-ballot wave is landing on America’s shores this November. Writing for the National Journal and his own election ratings and analysis site The Cook Political Report, Cook argued:
Two factors make me suspect that even if Clinton wins by a much larger margin than, say, Obama’s win over Romney four years ago, I don’t think the down-ballot implications would be that huge. In the House, there are fewer competitive districts than at any point in our lifetimes; between natural population sorting and gerrymandering, there just isn’t much elasticity in the House these days. In the Senate, the GOP majority is absolutely on the line; my guess it will end up 50-50, give or take a seat or two, but given voters’ doubts about Clinton, the “don’t give Hillary Clinton a blank check” argument may well be a politically potent one, and a lot of hold-their-noses Hillary voters may well look for a check and balance down-ballot.
Other publications are skeptical, too, noting that Senate incumbents are generally well over-performing Trump’s numbers in their states. In fact, Trump’s weakness could embolden Republican donors to abandon their presidential nominee and pour unprecedented resources into House and Senate races.
This is a plausible interpretation of history and current polling data, but it gives insufficient weight to four factors. First, it’s not clear that current congressional polling data has settled or fully factored in Trump’s potential drag. Unlike the presidential race, most congressional campaigns don’t get meaningful underway until after Labor Day, when congressional candidates’ more limited advertisement budgets begin to be deployed and when voters start focusing on races that get considerably less media attention. Moreover, many states only recently determined their congressional nominees. In fact, some states have still yet to hold their congressional primaries. Marco Rubio’s narrow polling lead in the Senate Florida race, for instance, is hard to evaluate given that both he and his likely challenger are in the midst of primary races that won’t be settled until August 30. Only recently have the generic ballot numbers for congressional races started to worsen for Republicans. Recent swing state polls now give Democrats an advantage in surveys asking whether voters plan to vote for a Democrat or a Republican for Congress for the first time this cycle.
The same dynamic presented itself in 2008. Sure, certain races in blue states with unpopular incumbents were already reliably showing large Democratic leads, such as in the New Hampshire Senate race pitting Jeanne Shaheen against John Sununu. But among the eight Senate seats Democrats gained in Obama’s first run for the White House, many Democratic challengers trailed early before grabbing leads in the fall that they would never relinquish. Jeff Merkley trailed Oregon Senator Gordon Smith in every major poll until late September. He won his race by over 3%. Kay Hagan trailed North Carolina Senator Elizabeth Dole in every poll before mid-August; after the conventions she nudged into the lead before winning handily. Finally, the Minnesota Senate race was back and forth all the way through the cycle before Al Franken was able to knock off Senator Norm Coleman by a nose.
The counter to this history is that the dynamics of the 2008 race were very different from 2016 in two important respects. First, by this time in 2008, it was very clear that outgoing President George W. Bush would define the race and place a considerable drag on the Republican nominee. According to Gallup, Bush’s favorable-unfavorable ratings in mid-August 2008 were an unsightly 33% favorable to 61% unfavorable. It is one thing for a senator to run away from an unpopular nominee; it’s quite another to avoid the deep unpopularity of a president they enabled. Second, unlike Hillary Clinton this time around, 2008 Barack Obama was remarkably popular in his own right. Favorable/unfavorable numbers from August 2008 polls had Obama anywhere from +7 to +31. Clinton is currently -11 (though improving since the conventions). But it’s at least worth noting that disparities in presidential and down-ballot polling at this stage of the race are not necessarily indicative of limited coattails.
Second, Cook fails to consider that Clinton could actually expand her already prodigious lead. As Nate Silver at FiveThiryEight recently pointed out, Clinton’s chances of winning by twelve points are roughly equivalent to the chances that Trump wins at all. Silver’s model currently assigns 35% odds that Clinton will win by double-digits. Another two to three points alone could be the difference in a host of House races that would otherwise stay Republican.
Third, Cook makes no more than a passing mention of redistricting and ideological sorting in support of his assertion that House races are more stable “than at any point in our lifetimes.” True, the redistricting process that took place after Republicans swept statehouses in the 2010 midterms has made things more difficult for Democrats. But difficult district maps and a polarized electorate existed in 2006 and 2008, when Democrats won thirty-one and then twenty-four House seats in back-to-back elections. Plenty of races where Republicans are accustomed to winning by 15% could be put in play with depressed Republican turnout and a strong number of cross-over votes at the presidential level.
Fourth and finally, Cook ignores the most important factor working in Democrats’ favor: the sheer number of seats Republicans have to defend in both the Senate and House in 2016. Consider the chart below, which lists the congressional results for each presidential election since 1964:
The “Dem. Win %” column lists the Democratic presidential candidate’s margin in the popular vote. The next column lists the number of Senate seats Democrats had to defend more than the number that Republicans did. The “Dem. Tot. House Seats Up” column lists the number of Democrats in the House of Representatives before that election (there are 435 members of the House and 218 needed for a majority). The other two columns list the net gains for Democrats in both chambers.
The most immediately apparent takeaway from this chart is how little a correlation there appears to be between a party’s winning percentage in the presidential race and the number of seats it picks up in Congress. At the same time, the number of seats a party has to defend clearly has an important effect. And when you combine the two factors together—a healthy presidential winning percentage and fewer seats to defend—that’s when the coattails begin to appear. In 1980, Republicans won by nearly ten points in the presidential election and had many fewer Senate and House seats to defend. They won twelve Senate seats and thirty-four House seats as a result. The reverse advantage appeared for Democrats in 2008. These are the exact circumstances present in 2016: Democrats have their lowest number of House members in decades and fourteen fewer Senate seats to defend.
Historian KC Johnson made a similar point in disputing an earlier Cook Political Report article by Jennifer Duffy discounting the possibility of a Democratic wave in 2016 through comparison to 1964, 1972, and 1984. Lyndon Johnson’s 1964 landslide may have failed to result in significantly larger congressional majorities, but that result was due to Democrats’ already strong performance in the 1958 midterms—where the party won a whopping fifteen seats—not the absence of a coattail effect. It was that large Senate class whose seats were at issue six years later. As Johnson notes, Democrats had significant vulnerabilities in 1964. They had vulnerable freshmen incumbents who won in difficult states in 1958 due to unique circumstances. They had open seats due to death and retirement. They had nominees emerging from difficult primaries. Meanwhile Republicans only had nine seats to defend, eight of which had senators running for reelection. That Democrats were able to win two more Senate seats under those conditions actually demonstrates a strong coattail effect, not a muted one.
So no, current polling showing Republican incumbents up even as Donald Trump trails by large margins does not mean that there will not be a down-ballot wave. Republicans have a large majority in the House and a disproportionate number of seats in the Senate to defend to go along with their deeply unpopular nominee. That combination will put seats previously thought to be unquestionably safe in play. Republicans have a right to be worried.