A friend said over the weekend, “The bird is in the oven. On Tuesday we will see what emerges—a goose or a buzzard. Probably a cross-breed of some sort.” That sums up pretty well the feeling among a lot of strategists and elected officials on both sides. Each side can envision or rationalize a favorable outcome but is also well aware of how this could go badly wrong. With so many really close races, a half-point or one-point swing either way can make an enormous difference in terms of seat counts and the overall national outcomes. That’s why throwing numbers around willy-nilly makes me so nervous.
For President Trump and his allies, the 2018 midterm elections amount to a political stress test of their double-down strategy. Rather than trying to expand his coalition beyond his 46 percent popular vote, this is about firing it up, hoping that the working-class, small-town, and rural-whites strategy that worked for him in 2016 can be replicated without Hillary Clinton in the equation. It is very plausible—actually it is likely—that it will work in the Senate, largely fought in red, conservative states, with essentially the same math that translated into 304 electoral votes for Trump (in both the Senate landscape and the Electoral College there is a thumb on the scale for less populated states). The test will be whether this strategy will work in House districts with large mixed suburban and urban constituencies, or at least work enough to keep GOP losses down to 22 seats or fewer, preserving the majority in that chamber, and keeping the number of governorships and state legislative chambers lost down in the lower single digits.
One finding Republicans have to like is on intensity. In the merged January-through-August NBC News/Wall Street Journal polls, Democrats had a 10-point lead on intensity, with 65 percent of Democrats calling their level of interest in the election a nine or 10 on a 10-point scale, to just 55 percent of Republicans. The gap narrowed to just 1 point in the latest, Nov. 1-3 survey, suggesting that Trump’s focus on the border wall, the caravan, and Kavanaugh has helped them a good bit. The question is whether it helped them only in red states and areas, or jacked up GOP turnout as well in purple suburbs where the House might be won or lost.
For Democrats, Tuesday will be a test of whether problems they encountered in 2016 were Clinton-related, or revealing of a larger problem in the party and its direction. One of the most-asked questions in recent weeks has been whether the polls are wrong, as some (particularly on the state level) were two years ago. Talking to polling experts, the surprising 2016 outcomes in Michigan, Pennsylvania, and Wisconsin were triggered by different things. In Michigan and Wisconsin, two states that were not widely considered close, Democrats lost largely because African-Americans did not turn out in expected numbers in Detroit and Milwaukee. That was true of some non-black Democrats in urban areas as well; some say that it was inattention from the Clinton campaign headquarters which caused that to happen in those two states. Meanwhile, in Pennsylvania, which always was a top priority for the Clinton campaign, Democrats got the numbers they needed out of Philadelphia, Pittsburgh, and surrounding suburban counties, but were swamped by unprecedented turnout levels in the area between the two cities and across the top of the state, reminding us of the James Carville description of the Keystone State: Philadelphia in the east, Pittsburgh in the west, and Alabama in between.
The relevance to this year is obvious. Outside of Florida and Georgia, where high-profile African-American candidates are in neck-and-neck gubernatorial races and sure to bring out a very large black vote, in the absence of Barack Obama, will African-American turnout be sufficiently high to avoid replication of Democrats’ 2016 Michigan and Wisconsin problems? Then there is the Pennsylvania side of the equation: Will Democrats be so thoroughly skunked in the red areas outside of the blue urban and purple suburban environs where Trump is problematic at best, if not toxic?
Similarly, will younger and Latino voters whose turnout levels were underwhelming in 2016 respond this year to reverse that outcome? There are some intriguing early numbers, particularly among those age 18-34, but one could have gone broke betting on young voters participating in midterm elections in the past. Or can fired up college-educated and suburban women, many of whom were unenthusiastic about Clinton, make up for shortcomings elsewhere?
Given the strong tendency of midterms as referenda on incumbent presidents, and the relevancy of presidential job approval to midterm outcomes, the variance in Trump’s approval ratings muddies the water. Using the oldest standard, when Gallup presidential approval has been at 46 percent or lower, the average losses are 40 seats in the House and seven seats in the Senate—though actually there aren’t seven Senate seats the GOP could lose Tuesday. When approval is 58 percent or higher, it’s been no net change in the House and a loss of two seats in the Senate. The Gallup Poll for the week through Sunday night put Trump’s approval at 40 percent for the second consecutive week, after it had been at 44 percent for the two previous weeks. But the NBC News/Wall Street Journal poll, conducted by Hart Research on the Democratic side and Public Opinion Strategies on the Republican side, puts Trump’s job approval rating at 46 percent approval (52 percent disapproval) among both likely and registered voters—1 point better than President Obama's before his disastrous 2010 midterm and 2 points above President Clinton’s before his equally calamitous 1994 midterm loss. The Oct. 29-Nov. 1 ABC News/Washington Post poll came in at 44 percent, and CNN’s Nov. 1-3 poll was a point above Gallup’s at 41 percent. Suffice it to say that the president’s numbers are well within the midterm-election danger zone, but it matters a lot where the president’s rating falls within that range.
Was 2016 a one-off election or a new norm, and will the two parties learn anything from Tuesday’s results? If Republicans focus primarily on the Senate and they do in fact expand their majority by a seat or two, will that mask the symptoms of real problems they have in more populated areas in general and suburban America in particular? Conversely, if Democrats win a House majority with say 30, 35, or more seats—turbocharged by a hyper-intense vote from suburban and college-educated women—will they ignore the nagging problems they have with white working-class voters and those who live out in small-town and rural America?
Both sides need to remember after tomorrow that overall midterm-election results have little predictive value about what will happen in the next presidential race. At the same time, if Democrats score as well as they might in the Michigan, Ohio, Pennsylvania, and Wisconsin gubernatorial and other state elections as well as in congressional races, will that suggest that whatever problems they had in those states in 2016 are not as applicable going forward? And what about Florida and its pivotal gubernatorial and senatorial races? Will their outcome give us some signs about the direction of a state that Obama won by a point in 2012 but Trump carried by the same margin in 2016?
The conventional wisdom is probably right that Democrats will pick up between 30 and 40 seats in the House, along with a half dozen governorships and state legislative chambers, while Republicans pick up a seat or two in the Senate—a wash at worst. But I confess to having some nagging notions that this understates the wave, that we are putting too much weight on geography and not enough on the national mood. We will find out soon enough.