Trump's low approval rating appears to put Democrats in an opportune position to retake the House, pick up seats in the Senate, pick up states in gubernatorial elections, and make gains in state legislatures in 2018. This is based on our analysis of this year's polling of congressional and gubernatorial races and their relation to Trump's approval rating.
We culled data collected from 15 polls which asked respondents which candidate they were most likely to vote for in a race between a Republican and Democrat and in which respondents were asked whether or not they approved Trump's job in office (question wording typically as follows: "Do you approve or disapprove of the way Donald Trump is handling his job as president?"). We found a staggering 87% correlation coefficient between the results of these two questions.
On average, GOP candidates outperform Trump's approval rating by only two points - and both track each other pretty closely, with an 87% correlation coefficient (R).
|Pollster||Election||Midpoint Date of Polling Fieldwork||Trump Net Approval Rating||Vote Margin||Shift from Approval to Margin|
|Quinnipiac University||New Jersey Governor||January 28, 2017||-19||D+16||R+3|
|Quinnipiac University||New Jersey Governor||March 11, 2017||-25||D+22||R+3|
|FOX 5 Atlanta/Opinion Savvy||Georgia 6th congressional district||March 23, 2017||+8.7||D+1||D+9.7|
|Gravis Marketing||Montana at-large congressional district||April 6, 2017||+8||R+12||R+4|
|Quinnipiac University||Virginia Governor||April 11, 2017||-21||D+11||R+10|
|FOX 5 Atlanta/Opinion Savvy||Georgia 6th congressional district||April 13, 2017||+8.5||D+2||D+10.5|
|Gravis Marketing||Montana at-large congressional district||April 27, 2017||+12||R+13||R+1|
|Quinnipiac University||New Jersey Governor||April 29, 2017||-21||D+25||D+4|
|Gravis Marketing||Montana at-large congressional district||May 3, 2017||+6||R+8||R+2|
|Gravis Marketing||Georgia 6th congressional district||May 9, 2017||+3||D+2||D+5|
|Gravis Marketing||South Carolina's 5th congressional district||May 20, 2017||+15||R+13||D+2|
|Gravis Marketing||Montana at-large congressional district||May 22, 2017||+8||R+14||R+6|
|Quinnipiac University||New Jersey Governor||June 9, 2017||-38||D+24||R+14|
|FOX 5 Atlanta/Opinion Savvy||Georgia 6th congressional district||June 15, 2017||+1.6||D+1||D+2.6|
|Monmouth University||Virginia Governor||July 21, 2017||-20||D+0||R+20|
The spread of the difference between how well a GOP candidate and Trump polls is a standard deviation of 8.3. Assuming a mean outperformance of 2 points, the 95% confidence interval (a range of values which correspond to 95% of possible outcomes) for how well Republican candidates do relative to Trump's popularity rating lies anywhere between a 15.6-point outperformance to an 11.7-point underperformance. This means that 95% of the time, a Republican candidate will poll a margin which lies within +15.6 to -11.7 points of Trump's current approval rating in the area in which they are running.
On a national scale, Trump's net approval rating is severely underwater, with FiveThirtyEight's polling aggregator placing its weighted average with likely and registered voters at -15 points (39.9% approve and 54.9% disapprove). This barely places Trump's national approval rating within our 95% confidence interval of possibilities, +15.6 to -11.7 points, of how much a GOP candidate's polled vote margin would exceed or fall short of Trump's approval rating where he or she is running. This means that using our aforementioned normal distribution model, we see that the probability of a GOP candidate winning in an area where Trump's net approval rating is the same as his national one (-15 points) is approximately only 5.8%
. This should obviously appear as a canary in the coal mine for Republicans, in the context of which the entire House of Representatives and a third of the Senate will have its seats up for re-election in a highly-competitive national environment in 2018. Additionally, gubernatorial elections will be held in 36 states.
There is only one poll in which a GOP candidate exceeds Trump's approval rating by 15 points or more (Trump's average net approval rating with U.S. voters is -15 points, per 538), and it is one from Monmouth University, of the Virginia Gubernatorial race to be held this year, in which GOP candidate Ed Gillespie's margin against Democrat Terry McAuliffe beats Trump's net approval rating by 20 points. That being said, this is only one outlier out of 15 polls - where, to reiterate, the average Republican candidate only outpaces Trump's approval rating by 2 points.
The formula for the trend line (linear least squares regression) pictured in the graph near the beginning of this page is y = 0.7118304522*x+0.54
, where y represents the predicted value of a GOP candidate's margin based on x, and x represents Trump's approval rating in the area in which the GOP candidate is running. Using this formula, based on an x value of Trump's national net approval rating of -15, a GOP candidate would be expected to lose by 10.1 points to a Democrat in an area representative of the nation as a whole. This is pretty much right in line with what generic ballot polls are forecasting for the outcome of the 2018 House race's popular vote, which RealClearPolitics and 538 pegs at a 9-point and 8-point loss for Republicans, respectively.
Some Republicans seemed to have taken their party’s wins in this year's competitive congressional races to mean that all is well in 2018. These races include the Georgia 6th congressional district election, Montana at-large congressional district election, and South Carolina's 5th congressional district election. In all of these districts in which a GOP candidate won in this year's house races, Trump had positive net approval ratings - while nationally, the picture is very different, with a double-digit negative approval rating.
In short; these districts liked Trump. On average, Trump had a net approval rating of +8.5 points in the Montana at-large congressional district, +5.5 points in Georgia's 6th congressional district, and +15 points in South Carolina's 5th congressional district. This means that on average, these red districts approve of Trump 24.7 points more than the overall U.S. population does. This lends credence to the idea the slim victories the GOP pulled off in these districts may not translate well to the broader picture of 2018.
Nationwide, Trump's approval rating is at an all-time-low since the beginning of presidential polling, and it appears that Trump's approval rating predicts the performance of Republican candidates. Unless Trump makes significant strides in his approval rating, which is altogether possible, this is obviously good news for Democrats come 2018. However, there are still geographic and structural barriers that may not make this an easy win for Democrats - Democrats would "need to win the House popular vote by about 8 percentage points to win half the House seats", according to 538's Harry Enten
. All signs currently show Democrats being just about that point, if not slightly past it - implying something around a 60-40 chance of Democrats retaking the House.
Appendum: This article was cross-posted to www.dailykos.com
There is an 87% correlation coefficient between a GOP candidate's vote margin in a poll and Trump's net approval rating. In polls so far, GOP candidates do about as well as Trump's approval rating, which seems to be their ceiling of support. With Trump's approval rating at a historic low nationwide, this could spell disaster for the GOP.