In political handicapping you often have to reconcile a conflict between macro and micro dynamics. The macro, or big-picture factors, focus on the sitting president, with his popularity highly relevant to the election outcome. The micro factors include seats and incumbents that are up for election, local political circumstances, and the demographics and voting patterns of individual states and districts. It is not at all uncommon to find macro factors favoring one party and micro factors favoring the other.

The 2018 midterms are a vivid example of the macro opposing the micro. History tells us that the president’s party almost always loses House seats, which has happened in 35 out of the 38 midterm elections (92 percent) since the end of the Civil War. In the Senate, the pattern is not quite as strong. Since 1913, when the 17th Amendment was adopted and the direct election of Senators began, the president’s party has lost seats in 19 out of 26 elections (73 percent). Exceptions have occurred: In 1998, a backlash against the impeachment of President Clinton helped Democrats gain House and Senate seats, and in 2002, 14 months after the 9/11 attacks, President George W. Bush’s Republican Party also picked up seats in both chambers.

When a president has job-approval ratings of 50 percent or higher, his party tends to keep its losses fairly low. But in six of the seven midterm elections since 1966, when presidential approval ratings hovered below 50 percent, his party has lost two dozen or more seats in the House, giving the opposition party a majority the next year. The lone exception was 2014 in President Obama’s second term. Democrats lost only 13 seats, but they had been all but destroyed in the 2010 midterms and hardly gained seats in 2012, so they had few competitive districts to lose in 2014.

President Trump’s approval ratings have averaged 40 percent, though they dipped to a worst-yet 34 percent approval and 61 percent disapproval in Gallup’s Friday, Saturday, and Sunday sampling. Unless some major national crisis occurs to cause a rebound in his popularity, the macro case for substantial GOP losses is pretty strong.

Talk about Democrats having problems because they don’t have a compelling message flies in the face of midterm-election history, with the exception of the 1998 impeachment backlash. Midterms are never about the party out of power; they’re about the party in power. Not having a clear message is problematic in presidential years, but not midterms. In short, a midterm election for a party with a president with job-approval ratings below 40 percent augur a pretty ugly Election Day in 2018.

But micro factors suggest a very different outcome. Cook Political Report House Editor David Wasserman made the case in an Aug. 7 article for Nate Silver’s FiveThirtyEight. While the Democratic Party may have long-term demographic trends working in their favor, Wasserman noted, the geography is very clearly working against them, in both the House and the Senate. Wasserman pointed out that even if Democrats in 2018 win every single House district and Senate race where Hillary Clinton either won or lost by less than 3 points, Democrats could come up short of a majority in the House and lose five Senate seats.

Many political analysts blame the Democrats’ problems on redistricting, and it is true that in the last remapping in 2011, after the Democrats’ disastrous gubernatorial and state legislative losses the previous year, Republicans were able to draw lines in many states to their benefit. In addition, because of population patterns and clustering of likeminded people in specific areas, the Democratic vote is highly concentrated in urban centers and in a few states like California and New York. This has hurt the party’s ability to win majorities in the House and Senate. Wasserman pointed out that Clinton’s margins in California and New York accounted for more than twice her popular-vote margin for the entire country.

Wasserman argued that you can calculate the level of partisan bias in House and Senate results. In 2008, for example, Barack Obama won the popular vote by 7.3 percentage points but won House seats by an average of just 4.4 percentage points. In 2016, Donald Trump lost the national popular vote by 2.1 percentage points but Republicans won the median House seat by 3.4 percentage points and the median Senate seat by 3.6 percentage points. Wasserman was quick to point out that this “doesn’t mean Democrats can’t win the House and Senate back—they won control of both chambers in 2006 despite a Republican bias that year, for example—but they’re starting from a truly historic geographic disadvantage, even with the political wind at their back.”

Exacerbating the challenge for Democrats in the Senate is the calendar. The class of Senate seats in play this year are those that were last up in two fantastic years for Democrats—2006, when President George W. Bush’s Gallup approval rating was 38 percent in the final preelection poll and the GOP lost six Senate seats, and 2012, when President Obama prevailed over Mitt Romney by a wider than expected margin helping Democrats gain two more Senate seats on top of the half dozen gained six years earlier. Thus Democrats enter 2018 with a class that is incredibly overexposed, as Wasserman noted. They are defending 25 of their 48 seats, while the GOP has just eight of their 52 up. Indeed Democrats have five incumbents up in states that Trump carried last year by 19 points or more, another five are up in states that he won by single digits. Just one Republican-held seat is in play in a state Clinton carried.

Cook Political Report senior editor and Senate specialist Jennifer Duffy is more skeptical of the more apocalyptic scenarios for Senate Democrats, believing that many of the Democrats up in the toughest states—Joe Manchin in West Virginia, Heidi Heitkamp in North Dakota, and Jon Tester in Montana, in particular—are no pushovers. She feels that it will be a challenge for Republicans to upend them when the political winds are in their faces. Duffy believes that the macro and micro forces may end up being something of a wash with competing factors largely cancelling each other out and, at least in the Senate, neither party is likely to see big gains or losses.

So will the 2018 midterm be a macro election that would likely feature a deeply unpopular president dragging down his party’s candidates? Or will it be a micro election in which Republican-leaning states and districts vote for GOP candidates, just as Democrat-tilting states and districts vote for their candidates, with the results that Democrats come up short again?

This story was originally published on on August 15, 2017

More from the Cook Political Report