
Election 2022 House Mississippi
President's Party Gaining in Midterm Polls. That's Not Normal.
ANALYSIS
Published Aug. 23, 2022 • Read time: 7m
For a few days last week, and for the first time this year, the Democratic Party inched ahead of the Republican Party in the RealClearPolitics Generic Congressional Vote average by 0.2%. As of Sunday, the GOP reclaimed the lead by the same, tiny amount. Nevertheless, the average has experienced a 4.6% shift toward the Democrats over the last four months, coinciding with a backlash against the overturning of Roe v. Wade, intensifying investigations into Donald Trump and his associates, lower gas prices, and a string of legislative victories for Democrats – both bipartisan and partisan.
Such a momentum shift toward the president’s party is unusual. Typically, over the course of a midterm election year, “the out-party gains in the polls and ultimately at the ballot box,” according to a 2010 political science research paper from three professors at different universities, based on data from the 16 prior midterm elections.
The RealClearPolitics averages in the 2010, 2014, and 2018 midterms largely bolster that analysis. In 2010, when Barack Obama held the presidency, Democrats did manage to gain three points on the Republicans between late April and early July and briefly hold a wisp of a lead. But by mid-July, as the conservative “Tea Party” revolt brewed, Republicans took back the lead for good, ending the campaign with a whopping 9.4% margin in the poll average en route to flipping control of the House.
Four years later, Democrats clung to a small lead for most of the period between early March and early September, but Republicans seized a lead by late September that was not relinquished, largely holding between two and four percentage points. On Election Day, Republicans retained the House and flipped the Senate.
In 2018, with Republican Donald Trump in the Oval Office, Democrats enjoyed a steady and sizeable generic congressional ballot lead all year. They held a 7.3% average margin by Election Day, more than enough to reclaim the House.
There are a couple of examples where the president’s party made late gains in the congressional generic ballot test: 1998 and 2002, which just happen to be the two most recent midterms when the president’s party bucked history and gained seats in the House.
The RCP archive of 2002 poll data doesn’t include the daily average. But using the raw data, I calculated a running average using every set of five consecutive polls. The calculations showed Democrats holding a small lead through much of the year, in temporary defiance of President George W. Bush’s stratospheric job approval numbers following the Sept. 11, 2001, terrorist attacks. But Republicans took a small lead in September 2002, Democrats briefly took it back in October, and then Republicans claimed a final lead of 2.4% at the top of November in advance of Election Day. The GOP picked up eight House seats to keep control, and two Senate seats to take control.
The data from Gallup’s generic congressional ballot test from 1998 – when Bill Clinton’s Democrats benefited from backlash against the Republicans’ autumnal pursuit of impeachment – has a zig-zag quality. Republicans led in April and August polls, but Democrats took a one-point lead in early September. That grew to a six-point Democratic lead soon after Republicans released the findings from the independent counsel’s investigation of Clinton’s extramarital affair. Republicans had a brief two-point edge in a late October poll, but the final Gallup poll gave Democrats a four-point advantage. That somewhat overstated the final result, as Republicans maintained control of the House with the help of a one-point margin in the overall popular vote. But the president’s party did pick up five House seats, a very rare midterm feat. (Before 1998, it hadn’t happened since 1934, and as of today, it hasn’t happened since 2002.)
So a late polling shift toward the president’s party is not only unusual, when it has happened, it has tracked with unusually strong outcomes for the president’s party. However, shifts in momentum after August have happened before, in both directions. That means Democrats cannot presume that their positive summer trend will automatically continue into the fall.
Beyond the unknown of future news developments, Democrats don’t know how big a lead in the generic ballot poll average they need to keep the House. Conventional wisdom, as articulated recently by CNN polling editor Ariel Edwards-Levy, is that “Democrats typically need a healthy lead on that metric to translate into winning numbers in the House ... partly because of redistricting, and partly because Democratic voters tend to live in more densely populated areas.”
It happens to be true that Democrats had big generic ballot poll leads in the final RCP average the last four times they won the House: 11.5 points in 2006, 9 points in 2008, 7.3 points in 2018 and 6.8 points in 2020. But in the first three of those four wins, the Democratic share of the House popular vote and, most importantly, House seats, was not narrow. Democrats didn’t need such big poll margins to reach the magic number of 218 House seats.
Only in 2020 did the Democrats win a majority of House seats with just a single-digit margin. And that year, the Democratic margin of victory in the House popular vote (3.1 %) trailed their final margin in the RCP average. That one example does suggest Democrats need a cushion in the polls. But it is not the case that the Democrats’ House popular vote performance always trails their poll performance. In the 10 House elections for which RCP produced a generic congressional ballot average, Democrats outperformed the poll average four times.
Furthermore, in the last two House elections, the share of Democrats’ House seats roughly matched their share of the national popular vote. In 2018, Democratic candidates won 53.4% of the votes and 54% of the seats. In 2020, they won 50.8% of the votes and 51% of the seats. In other words, Democrats did not need to win a disproportionate number of votes to eke out a narrow win. And this year’s post-census redistricting process appears not to have wildly skewed the map toward the Republicans. Dave Wasserman of The Cook Political Report estimates that Republicans have gained just three House seats from redistricting, which helps them close the five-seat gap from the 2020 election, but doesn’t close it completely. (Also of note: Michael Li and Chris Leaverton of the Brennan Center note that a majority of the districts on the new map, 227 out of 435, are districts won by Joe Biden in 2020.)
In other words, it’s possible for Democrats to hang on to the House with only a narrow House edge in the popular vote – just as it’s possible that a narrow poll edge will translate into such a vote.
Of course, possible is not probable. Popular vote underperformance of around two to four points, relative to the polls, was the case for Democrats in six of the last 10 House elections. Democrats would surely feel much more confident with another seven-point poll edge, like the last two elections, than being within the margin of error.
Democrats have already done something rarely seen in a midterm year: They gained polling ground over the summer while holding the presidency. If the election was today, those gains probably wouldn’t be enough for Democrats to keep the House. But if that momentum continues over the next two months, watch out.
RCP Today
Top stories on RealClearPolitics
The Takeaway
Polling intelligence and analysis delivered to your inbox
World News
Featuring the latest world news
Policy
Featuring the latest policy news
Support Our Mission
As part of RealClearPolitics, authors aim to provide balanced, non-partisan analysis by presenting top reporting and commentary from all sides of the political spectrum.
Trusted by policymakers and citizens alike since 2000






