On November 8, 2022, the United States will hold midterm elections for both the House of Representatives and the Senate. All 435 House seats are up for reelection and 34 Senate seats are up as well. The outcome of this election will determine the direction of the country for the next several years. If the Republicans take the House, the Biden Agenda will linger in the depths of the back burner for at least two more years. And if the Republicans retain control of the Senate, they will only have to face a Presidential veto, which they may have the ability to over ride.
Many pundits are predicting that the Democrats will loose control of both the House and the Senate, based mainly on Biden’s shrinking Job Approval ratings from a number of creditable polling firms. Most political scientist fall in one or two categories: the popularity/economy theories and believe that midterms are a referendum on the current President’s performance which is often represented by the his Job Approval rating. In other words, the lower it goes, increases the number seats loss.
Other researchers believe the economy is the main factor in whether voters either support or oppose the President’s party at midterms. If the economy is good, they vote to support the President’s Party. If it’s bad, they vote for members of the opposing Party.
In this post, I’m going to test these common hypotheses to see if the President’s Job Approval actually predicts how many House and Senate seats the Democrat’s will win or loose in 2022. And if so, what are the likely changes in the House and Senate. I’m also going to also measure the impact of the economy on the midterms as well and in combination with the President’s Job Approval rating.
Since 1962, when John F. Kennedy faced his only mid-term election, the number of House Seats lost by the his party in the mid-term elections was 23 and 3 for the Senate. (Democrats lost 27 vs. Republican’s 20 House seats). Since the 1962 midterm election, only two Presidents managed to gain seats in the House. Bill Clinton in 1998 gained 4 House seats and George W. Bush in 1992 picked up 8 seats in 2002. In the Senate, only four Presidents managed to add additional seats: John Kennedy (4), Nixon (2), G.W. Bush (2) and surprisingly, Donald Trump gained 2. In other words, the odds that a sitting President will lose seats in the House and Senate are pretty good, regardless of party.
Using Multiple Regression, I have tested the importance of each economic measure in each midterm election since 1962, that researchers commonly use to measure the health of the economy: G.D.P. Growth, Unemployment, Inflation and Consumer Sentiment for each midterm year and, of course, the sitting President’s Job Approval rating on our dependent variable: the number of House seats won or loss. Below are the regression coefficients, but only one matters for our purposes here: the Significance of each variable under the title Sig. (significance)
Model
|
Unstandardized Coefficients | Standardized Coefficients | t | Sig. | ||||
| B | Std. Error | Beta | ||||||
| 1 | (Constant) | -101.211 | 32.718 | -3.093 | .013 | |||
| CONSUM.SENT | .051 | .235 | .053 | .218 | .832 | |||
| UNEMPLOYMENT.RATE | -1.464 | 2.746 | -.124 | -.533 | .607 | |||
| INFLATION | -1.161 | 1.515 | -.169 | -.766 | .463 | |||
| GDP.GROWTH | -1.453 | 1.824 | -.176 | -.797 | .446 | |||
| JOB.APPROVAL.PERCENT | 1.700 | .559 | .711 | 3.042 | .014 | |||
Table 1: REGRESSION MODEL COEFFICIENTS
The significance (sig.) of each variable tells if it is contributing anything to the dependent variable (number of house seats won/loss), while controlling for the effects of the other variables. The only significant variable is the President’s Job Approval rating (.05 or less represents a significant level). Simply put, all the economic variables had no impact on the number of house seats won or loss. The only variable that matters is the President’s Job Approval rating (ignore the Constant, for purposes here it is meaningless.)
This is surprising since many researchers believe that the economy is the driving force as to whether a Presidential incumbent’s approval rating is good or bad, but this equation adds some doubt to this common belief. When I make the Presidential Job Approval rating the dependent variable, these economic variables are not significant and thus don’t contribute anything to the President’s popularity.
Similarly, a recent academic article by Economic Professor David L. Kleykamp, (The Economy and Presidential Approval) also found that as well: “all of the common macroeconomic variables discussed so often in the news are of little or no importance in determining a president’s job approval rating.”) Consequently, I’m left with the conclusion that the only variable that effects Joe Biden’s success or failure at midterm, is the final job approval rating before the elections.
Can we predict the number of won or loss house and Senate seats the Democrats realize on November 8, 2022? The simple answer is yes, but the timing is the problem. A definitive answer will only occur when the final Job Approval rating before the midterm election is released. But we can estimate the number of won or loss seats today, using the equation’s unstandardized coefficients. In other words, we can estimate the the number of House seats won or loss based the current approval ratings as if the election occurred today. Our equation is simple:
(approve rating –disapprove rating ) x unstandardized rating = house seats won/loss.
Let’s pretend the midterm election occurred today and Quinnipiac has just to released its final Job Approval Poll. It shows that 37% approve and 52% disapprove of Biden’s Job. That equates to a negative difference of -15 (37 – 52= -15). Since our equation’s unstandardized rating is 1.7, we multiply it x (-15) = -25.5 or 25 House seats lost.
Below is a chart showing possible (hypothetical) midterm outcomes and how it would effect the number of House and Senate seats. Each net 2% increase in Biden’s Job Approval, equals to an increase of 3.4 House Seats. Below is chart that shows different hypothetical midterm outcomes.
| JOB DISAPPROVAL | JOB APPROVAL | Net Diff | HOUSE Won/Loss | SENATE Won/Loss |
|
42% |
56% |
-14 | -24 | -3 |
| 47% | 48% | +1 | +2 | 0-1 |
| 48% | 52% | 4 | +7 | +1 |
| 42% | 56% | 14 | +24 | +4 |
Table 1
Table 1 shows that for the Democrats to retain their slim majority they need a minimum Job Approval rating of 48% and a disapproval rating of 47%. And If the approval Rating reaches 52% (net 4%), they gain 7 House seats and 1 Senate seat. The big win for the Democrats is when the Approval rating is 56% and a Disapproval of 42%, which increases their House seats by 24 and Senate seats by 4.
At this writing, Biden’s average Job Approval rating according to 538 polls, is 43% Approve and 52% Disapprove, for a net difference of -9%. Using our formula, this means if the election were held today, the Democrats would lose 15 House seats and 2 Senate seats.
All this shows that Joe Biden’s Job Approval rating has to reach at least reach 48% (and a disapproval rating of 47%.) Anything less would mean the Democrats would lose the House. There is still plenty of time for Biden to recover his Approval rating As we get closer to the Midterms, I’ll update the estimated House and Senate seat counts. But in the meantime, whether you are a Democrat or Republican, pay attention to Biden’s Job Approval to see how your team is doing…
