Elections

How Will the Senate Races Break?

Incumbent Senator Ted Cruz (R, Texas) speaks to supporters inside Schobels’ Restaurant in Columbus, Texas, September 15, 2018. Cruz’s Democratic challenger, Beto O’Rourke, has raised roughly $38 million in just the last quarter. (Sergio Flores/Reuters )
A detailed look at the numbers.

How will the Senate races break? We have less public polling to go on than in recent years, so answering that question is harder than ever. But the news is more optimistic for Republicans than it was a month ago.

 

Waves and Breakers

Four years ago, I projected in mid September that if “historical patterns hold in 2014, we would . . . expect Republicans to win all the races in which they currently lead plus two to four races in which they are currently behind, netting a gain of 8 to 10 Senate seats.” This was far from the consensus at the time, but in the end, Republicans picked up nine seats.

As I noted at the time, drawing on Sean Trende’s work on the relationship between presidential approval ratings and Senate races, the most important factor in Senate races is the state’s political climate, which typically is tied to the national political climate. That’s not the only factor, of course; others, such as incumbency and candidate quality, also matter. But going back over the past two decades, most Senate elections have been held in at least mild “wave” conditions: The president’s approval rating and the generic congressional ballot both pointed in the same direction. And the results showed a strong pattern: Looking at the RealClearPolitics poll averages, candidates with the wave at their back and a polling lead or tie in a close race in mid September won nine out of twelve races between 2002 and 2012, whereas candidates running against the wave with a lead or tie in a close race in mid September lost 21 out of 34 Senate races. And of the ten races where the wave party overcame at least a 2.5-point poll deficit, six were against incumbents, showing that the wave environment was stronger than the entrenched value of incumbency.

That trend was even stronger in 2014. My method, tracking the 2014 Senate and governor’s races, was to follow the “breakers” — the undecided voters who gradually made up their minds or decided to stay home. Two Republicans saw the bottom drop out of weak campaigns (in Michigan and Oregon) and one Democrat actually lost support down the stretch (Mark Warner in the thinly polled Virginia race), but in all the others, you could approximate a measure of how late deciders contributed the final outcome by looking at what percentage of the remaining undecideds ended up on each side — and of the 15 races where both candidates were below 50 percent in mid September, Republicans won ten of the 15, and their final vote was consistent with getting at least 60 percent of the undecided vote in eight of their ten wins plus New Hampshire and Virginia.

But then, in 2016, something funny happened: Starting in March, Barack Obama’s job approval was in the black for the first time since 2012, and by mid September he was over 50 percent approval, around +4. The Democrats led in the generic ballot, too: by +2 in mid September and as much as +6 in mid October. Based on the RCP poll averages, the mid-September polls projected a D+3 outcome (Democrats gain seats in Illinois, Wisconsin, Pennsylvania, and Indiana, and lose a seat in Nevada). But in the end, the outcome got worse for Democrats: They ended up at D+2, gaining just one of the four races they led in mid September (Illinois, which had not been close) plus New Hampshire and reclaiming their lead in Nevada. But the Wisconsin, Pennsylvania, and Indiana races got away from them, and so did the seesaw race in North Carolina.

Did the 2016 Senate results buck a Democratic wave? Not really. True, Democrats won the national popular vote for president and picked up a net gain of six seats in the House. But their supporters were too geographically concentrated, with Hillary Clinton winning a popular majority in just 13 states and Republicans’ 241–194 House majority outstripping their more modest 49.1 percent–to–48 percent victory in the House vote. In the end, even though Republican Senate candidates ran ahead of the presidential ticket and often drew on different supporters than Donald Trump did, the Senate outcomes paralleled the presidential race: Every one of the 34 Senate races was won by the party that carried that state in the presidential race, the first time this has happened since senators first became popularly elected in 1913.

 

State of the Senate

That leaves us with our first dilemma in projecting 2018: Nationally, this should be a Democratic wave year, at least a modest one. President Trump’s job-approval rating on RCP has been underwater since a few weeks after he took office; it was 12.5 points in the red on September 15 (40.9 approval, 53.4 disapproval), and at this writing stands a little under ten points down, with his approval rating having stabilized again around 43 percent, roughly where he has been for most of the time since early May. The generic ballot has likewise shown a Democratic lead all year; it stood at D+8.3 on September 15, and at this writing is D+7.3. That is, both measures have tightened a bit in the GOP’s favor in the past month, but not dramatically enough to suggest anything but a blue wave of at least some size.

But as in 2016, the Senate is not a national election but a series of state-by-state elections with very different electorates. And this year has an unusually red map — by some measures, it’s the most hostile map any party has ever faced in a Senate election. Only nine Republican-held seats are up for grabs (with the incumbents seeking to hold six of them), and just one of the nine (Dean Heller in Nevada) is in a state Democrats have carried in a national election more recently than 1996. By contrast, the Democrats are defending 26 seats, including five in states where Donald Trump won 55 percent to 68 percent of the vote (West Virginia, North Dakota, Indiana, Missouri, and Montana), five others in states Trump carried narrowly in 2016 (Florida, Ohio, Wisconsin, Pennsylvania, and Michigan), four in states that have recently been competitive (Virginia, Maine, and two in Minnesota), plus one occupied by an incumbent who narrowly escaped going to federal prison a year ago (New Jersey). On the upside, Democrats are running incumbents in all 26 races.

Each side has recruited one truly top-shelf challenger: Republican two-term incumbent governor Rick Scott in Florida, and Democratic former two-term governor Phil Bredesen in Tennessee. Republicans are also running the most popular candidate imaginable for the open seat in Utah, Mitt Romney, who won 72.6 percent of the vote there in the 2012 election (only Ronald Reagan in 1980 and 1984 topped that since 1900). There is also one race with a significant enough third-party candidate to scramble the picture: former two-term Republican governor and two-time Libertarian presidential candidate Gary Johnson is running as a Libertarian in New Mexico.

So the first question we should ask is what the environment in those states looks like right now. State-by-state polling for presidential approval tends to be spotty, and for the generic ballot just about non-existent, but as at least a rough proxy we can look at Morning Consult’s monthly tracking poll, which as of September showed Trump’s presidency in very different light depending on the state:

Among the seats Republicans are defending, four are in states where Trump is still wildly popular (including two contested races, Tennessee and the Mississippi special election); two are in states (Texas and Nebraska) where he’s still got at least 50 percent support; and one is in a state (Utah) where Trump is unpopular but the state is too Republican, and Romney too popular, for that to matter. That leaves just two states where Trump’s approval looks anything like the national average, Arizona and Nevada, and even in Arizona, Trump at −3 (47–50) is not that big a drag.

By contrast, the five states that were deep red for Trump in 2016 now range from red to very red, and Trump is above water in Florida and not that far down in Ohio and, surprisingly, Virginia. But on the whole, his support in the Midwest has taken a dive out of proportion to the rest of the country: he’s at −8 in Pennsylvania, −13 in Michigan, −16 in Wisconsin and Minnesota.

So, when we do our first cut of the Senate races that you’d want to watch, as of the polls through October 15, you get this:

I’m not even showing Utah, where Romney is running away with the race, or New Mexico, where Gary Johnson may hold incumbent Democrat Martin Heinrich below 50 percent but has so split the opposition that Heinrich is in no danger. The first thing to notice here is the right end of the chart. Virginia, where Tim Kaine may always have been unbeatable, is the only race Republicans have really thrown away this year with a disastrous nominee (Corey Stewart could challenge Jim Gilmore’s 33.7 percent in 2008 for the worst showing by a Republican in a major race in Virginia in a quarter-century, and deservedly so). But outside of red Indiana, the Midwest is basically written off at this stage. Debbie Stabenow, Tammy Baldwin, Bob Casey, and Sherrod Brown are all above 50 percent (in Brown’s case, despite Trump remaining fairly popular in his state), and Tina Smith (Al Franken’s replacement) looks to hold a safe lead over state senator Karin Housley in Minnesota.

You might suspect that Michigan is not done just yet; John James has closed the gap to single digits in the most recent polling, and the polls got Michigan catastrophically wrong in both the Democratic primary and the general election in 2016. Maybe that’s the case, but there’s not that much we can do to quantify the possibility of another big miss, and I’d be more inclined to say James was still in the hunt if he were down 6 instead of 9 in his best poll, and 12 instead of 23 in his worst.

Taking those races off the table, we have eleven contested races that have been polled:

I say “that have been polled” because RCP, at least, does not view any of the polls in the Mississippi Senate special election as credible enough to count; we’ll come back to Mississippi later.

Bear in mind that, assuming no upset in Mississippi, Republicans need to win three of these eleven races to keep the Senate at least 50/50 (in which case Mike Pence would break ties to retain control of the chamber), four to keep their current 51-seat majority, and five to actually expand their majority. At first glance, this divides the key races into three groups. Three races — those in North Dakota, Texas, and Tennessee — look pretty solidly Republican. Four races — in New Jersey, West Virginia, Montana, and Florida — look pretty good for the Democrats. Four others are still practically toss-ups (in Nevada, Missouri, Arizona, and Indiana), with only Dean Heller in Nevada having clawed his way as far as 46 percent of the vote.

But the percentages can be misleading in a couple of ways. Let’s break down the polling we have, with Trump’s approval rating in each state for perspective on the state environment. In the seven races with no polling at all, I just plugged in the president’s approval rating as a proxy, but that undoubtedly understates how safe incumbents such as Deb Fischer and Maria Cantwell are:

The “Break” column here indicates what percentage of the undecided vote would need to break the Republican’s way to get to 50 percent (or, in the case of New Mexico and Maryland, 50 percent of the non-third-party vote). Mitt Romney, Kevin Cramer, and Ted Cruz, having hit 50 percent already, only need to avoid backsliding. The usefulness of this measure, rather than just a raw “who leads by how many points” top line, is that it takes account of the fact that races like those in Indiana and West Virginia are showing a lot more voters still up for grabs than, say, those in Tennessee or Montana. Measured in percentage points, Patrick Morrisey is further behind than Bob Hugin, but Hugin has a steeper hill to climb because his opponent, Bob Menendez, is closer to 50.

The second thing to watch is how much public polling we actually have, and how fresh it is. The “Last Poll” date listed on the chart is the date the most recent poll entered the field, not when it concluded. RCP is slightly more aggressive when it comes to excluding partisan and Internet polling than some other sites are, but no matter where you look today, we are seriously hurting for polls. Only six of the 35 Senate races have at least three polls that were in the field on or after September 27. I picked that date because it was the day of the televised Kavanaugh–Ford hearing, and whether or not that was the reason, we’ve seen a distinct shift toward Republicans across a number of polls (and not just in the Senate) since, calling into question the reliability of earlier polls. But in races where we have only one or two new polls, we have less reason to be confident that we’re not just falling for a bad poll because it’s all we have.

The six races with fairly extensive fresh polling are Texas, Tennessee, Nevada, Missouri, Arizona, and New Jersey. In blue New Jersey, Bob Hugin looks like a replay of Republican failures to close the deal in 2002 and especially 2006, when Tom Kean Jr. ran a strong campaign but faded down the stretch in a hostile environment. In Missouri, Josh Hawley looks to be ahead by a nose but has yet to really pick up ground. But the other four races have all shifted hard in favor of the Republican, in Martha McSally’s case closing to something like a dead heat amid a blizzard of fresh opposition research about Kirsten Sinema’s hard-left early 2000s background.

By contrast, we have only two recent polls in hotly contested Florida, just one in Indiana, and none in Montana or West Virginia. That means we actually know a lot less about what is happening in these races than the numbers indicate, even if we trust the polling after the final polls missed three Senate races (Wisconsin, Pennsylvania, and the razor-close race in New Hampshire) in 2016.

The failure to do any reputable polling in Mississippi is also puzzling. Mississippi uses a “jungle primary” system, so unless a candidate breaks 50 percent on Election Day, there will be a runoff three weeks later on November 27. What limited, sketchy polls we have suggest that the appointed incumbent, Cindy Hyde-Smith, is probably running ahead of the leading Democrat, former Clinton agriculture secretary Mike Espy (whom you may remember from his indictment by an independent counsel in the late 1990s for taking a bunch of petty graft). An April Mason-Dixon poll had Hyde-Smith up 46–34 and Chris McDaniel down 42–40 in a hypothetical runoff with Espy, and an October 4 NBC News/Survey Monkey Internet poll showed Espy at 25 percent, Hyde-Smith at 24 percent, McDaniel at 19 percent, and 27 percent undecided — tentative good news for Republicans worried that Espy might hit 50 percent on Election Day against divided opposition but not entirely reassuring news about Hyde-Smith’s ability to hold off McDaniel, who could threaten a repeat of the Roy Moore and Corey Stewart fiascos.

Now, if we look at the races that have actually been polled at least once both before and after mid September, do we see movement that suggests a wave?

Not a national blue or red wave, no. Outside of deep-blue Delaware and deep-red Utah, no race has shifted more than four points into the Democratic column, and only North Dakota and Tennessee have banked further than that into the Republican column. Overall, six races have shifted at least two points bluer, and one of those is the lightly polled, many-undecided Indiana race; six have shifted at least two points redder, but that includes the opponents of Elizabeth Warren and Kirsten Gillibrand enjoying the proverbial “dead cat bounce” without any real chance to win (similar to why Romney’s opponent has surged into the high 20s). But while this is overall an optimistic picture for the GOP, unless and until we see more polls showing that Florida, Indiana, Montana, West Virginia, or Missouri are enjoying movement to the Republicans, it’s not just hard to generalize about a positive trend in the Senate; it’s also hard to be too confident that the movement we have seen in the past month will continue. And of course, the polls could just be off a handful of points in either direction, affecting our perception of any number of close races.

The most interesting races of this cycle are the ones with “survivor” candidates — men and women who have shown, in some cases repeatedly, the ability to win tough races in which they were expected to be vulnerable. Examples this year include Democratic senators Joe Manchin, Heidi Heitkamp, Claire McCaskill, and Jon Tester, Republican senator Dean Heller, Democratic challenger Phil Bredesen, and, over in the governor’s races, Scott Walker. There are also many such races in the House. But maybe the most compelling race this year is the Florida Senate race, with both Senator Bill Nelson and Governor Rick Scott having outlived many predictions of their demise (in 2014, Scott’s critics were gloating that he had lost as late as the early returns on Election Night), but both having customarily won with the national wind at their backs (Scott in 2014 and 2010, Nelson in 2012 and 2006).

The final wild card, of course, is that many of these races are happening alongside contested races for governorships and the House. Republicans have opened up what look like blowout margins now in the Texas, Arizona, and Tennessee governor’s races, while Democrats hold a more modest lead in Florida, and Nevada is basically a dead heat.

If pressed to make a prediction right now, I’d say the Republicans have a very strong hand to hold on to 51 seats in the Senate, picking up North Dakota and maybe Missouri while having a better than 50/50 shot to hold at least one of Nevada and Arizona and increasingly strong odds of retaining Texas and Tennessee. There remains a chance of a Democratic takeover, but it has faded of late, and something will need to interrupt the latest trend to reverse it. That said, Republicans will need fresher good news from Indiana, Montana, Florida, or West Virginia before they start giving real credence to their chances of adding more than one seat to their majority.

Dan McLaughlin — Dan McLaughlin is an attorney practicing securities and commercial litigation in New York City, and a contributing columnist at National Review Online.

Most Popular

Politics & Policy

Kat Timpf Chased Out of Brooklyn Bar

Fox News personality and National Review contributor Kat Timpf was forced to leave a bar in Brooklyn over the weekend after a woman she had never met became enraged upon learning she worked in conservative media. Timpf, who has twice previously been harassed while socializing in New York City, first described ... Read More
Film & TV

The Dan Crenshaw Moment

Given the spirit of our times, things could have gone so differently. On November 3, when Saturday Night Live comic Pete Davidson mocked Texas Republican Dan Crenshaw’s eye patch, saying he looked like a “hit man in a porno movie” — then adding, “I know he lost his eye in war or whatever” — it was a ... Read More
U.S.

The Present American Revolution

The revolution of 1776 sought to turn a colony of Great Britain into a new independent republic based on constitutionally protected freedom. It succeeded with the creation of the United States. The failed revolution of 1861, by a slave-owning South declaring its independence from the Union, sought to bifurcate ... Read More
Elections

Fire Brenda Snipes

Brenda Snipes, the supervisor of elections in Florida’s Broward County, does not deserve to be within a thousand miles of any election office anywhere in these United States. She should be fired at the earliest possible opportunity. Snipes has held her position since 2003, in which year her predecessor, ... Read More
Elections

Florida’s Shame, and Ours

Conspiracy theories are bad for civic life. So are conspiracies. I wonder if there is one mentally normal adult walking these fruited plains -- even the most craven, abject, brain-dead partisan Democrat -- who believes that what has been going on in Broward County, Fla., is anything other than a brazen ... Read More
World

How Immigration Changes Britain

Almost nothing is discussed as badly in America or Europe as the subject of immigration. And one reason is that it remains almost impossible to have any sensible or rational public discussion of its consequences. Or rather it is eminently possible to have a discussion about the upsides (“diversity,” talent, ... Read More