The Cook Political Report has a good (with some caveats explored below) piece on why the 2022 mid-term elections are likely good news for Republicans: Why 2022 Rhymes With the Previous Four Midterms. The headline is a reference to the alleged Mark Twain bon mot, “History never repeats itself but it rhymes.”
It has been known since the results of the 2020 elections were tallied that the likelihood was that the Democrats would lose control of Congress in the mid-terms. It is a long-standing pattern in American politics. A major reason for this, as we saw in the 2021 Virginia governor’s race, is that the electorate shifts in non-presidential elections in a way that tends to help the party that does not control the White House.
The piece notes four factors:
- One Party in Power
- President Biden’s Low Job Approval Ratings
- Enthusiam Gap
- Independent Voters
In regards to the first time, the trends are pretty clear:
The last president to hold onto both the House and Senate majorities post-midterm was Jimmy Carter in 1978.
[…]
The last three midterm elections which featured one-party control of the White House, House and Senate were 2006, 2010, and 2018. In all three cases, the president’s party lost the House. In 2010, Democrats lost seats but managed to hold the Senate.
Along these lines, the piece references a Pew study that notes:
Although a single party in charge in Washington is common at the beginning of a new president’s term, there has only been one presidency since 1969 where control has lasted beyond the following midterm election. That was during Democratic President Jimmy Carter’s one term in office, when Democrats retained leadership of the House and Senate in both the 95th and the 96th Congress (1977-1978 and 1979-1980).
Pew provides this useful graphic:
The list is really all linked to the same issue, in my view: once a president is elected, support will wane. An electorate selects a given president for a myriad of reasons. By definition, a given president will fail to deliver on a large number of those reasons. For one thing, voters (irrationally) expect a new president to come into office and change things, well, immediately. This never happens, at least not at a level to make most people happy. So there is a built-in strike one. Then, any president has to rely on Congress to enact major policy changes. Under ideal conditions, this is possible (which would include things like unified caucuses in both chambers and a little thing we like to call a filibuster-proof majority in the Senate), but more likely the tall orders linked to campaign promises combined with a very short window of opportunity makes making people happy unlikely. And, more important than any of that, are conditions largely outside the control of a given president (e.g., things like oil prices, pandemics, and the general state of the economy not to mention global strife). While, yes, policy matters for such things, it matters almost certainly less than people think, and even when it does matter, it sometimes takes years for its effects to come to pass.
To put it more succinctly: even with perfect conditions in Congress and in the world, the actual impact of policy is unlikely to be seen in the first few months of an administration, yet people start deciding if they approve of the president pretty much by late afternoon of Inauguration Day.
This is especially true since presidential approval ratings are essentially a proxy for “how are things going?” and not really an especially great metric of performance (made more complicated by the fact that partisans automatically dislike the president of the opposite party, not to mention that most people really don’t know how governing works).
As such, turn-out and results in a mid-term are driven by a climate that is automatically unfavorable to the sitting president. Further, the out-party is automatically more prone to show up to vote (back to the CPR piece):
there’s a fundamental truth that’s been very consistent over the years: Angry people vote, and complacent or disappointed people don’t.
This, more than anything else, is why the ‘out’ party has an advantage in midterm elections. Their voters are confronted with the consequences of losing the last election every day. That keeps them frustrated, angry and engaged (think, #resistance or #letsgobrandon movements). It’s much harder for the winning side to keep their voters engaged. This is especially true for someone like Biden, whose appeal to many voters wasn’t as much who he was as who he wasn’t.
This point is elaborated on in detail in the piece, to which I refer anyone interested. This is also true of the issue of independents, which is really just a subset of enthusiasm (at least as linked to turnout) and approval ratings.
Let me note that while a negative outcome for the president’s party is not guaranteed, it is nonetheless highly likely. Moreover, I would note that the nature of this pattern points to structural aspects of our political system far more than what a lot of “analysis” is going to say in the coming year (i.e, that it is about messaging or specific policies). There will be a lot of “If only Democrats had done X” or “Why did Democrats say (or not say) Y?” There will be plenty of “Dems in disarray” pieces. And while, sure, message/policy success are factors, they don’t matter as much as other variables such as the partisan makeup of districts.
One major thing the piece misses is the lack of competitive seats to begin with. There is, in my view, an ongoing problem with most discussions of American politics: the assumption that each election starts off with all competitors equally situated at the starting line and then the outcome of the competition is predicated on which party has the best message, the best candidates, and/or “does politics” better. The reality remains that the results in most districts (whether House districts or states as it pertains to the Senate) are largely knowable today. We don’t need to know anything about the quality of candidates or of campaigning to know which party will hold the seat. This needs to be discussed more than it is because the dominant narrative (even from people who would acknowledge this if the issue is raised) is such that it creates the illusion that electoral outcomes are about how well the parties compete, not other factors.
Let me stress: for that to be true (competition being more about policy than structure), we would need a wholly different system that actually allowed voter preferences to translate into seats. That is, we need some form of proportional representation.
Instead, our elections are less about the general preferences of voters and far more about how lines on the map dictate outcomes. This leads to a lot of non-competitive races that simply are not a function of message v. message.
Here’s the Senate forecast at the moment, as summarized by Ballotpedia:
Note out of 34 contests, a maximum of six are truly competitive (and Larry Sabato only has four that as competitive) with another four (at best) that might be non-blowouts. Note, of course, especially as it pertains to the “message” narrative, that only 1/3rd of the Senate is up for re-election at a given moment. This means that the post-election make-up of the Senate cannot be understood to be a function of how well the sitting president nor the current party in power is doing because 2/3rds of the chamber was elected under a different president and a differently constituted congress.
It is not possible to have a complete assessment of House races at this point, as not all the maps are finished, but the Cook Report currently has a grand total of 14 toss-up seats, with another 17 in the “leans” category. (Of course, the fact that the maps are needed for this discussion underscores the problem of single-seat districts: the lines are more important than the voters.
It also cannot be forgotten that the way our elections are scheduled (presidency every four years, but House and 1/3 of the Senate every two) is that we create very different electorates for presidential v. mid-term elections. This fact is simply not fully appreciated/accounted for in these conversations. Presidential elections draw more voters and create a different electorate (partly for reasons noted above) than do mid-terms. Regardless of anything else, this makes the narrative that elections are primarily about competing messaging miss the mark.
Yes, there is some influence of message, policy success, and candidate quality that matter, but not as much as it is made to sound in a typical discussion of American politics. If you ask two different sets of people their opinions, even if there is some commonality between those groups, you will very likely get different answers.
Indeed, it is a strange design to elect a president, and then give them less than two years to perform before deciding to give people a chance to revoke control of the legislature (especially given the electorate churn that I noted). It would make more sense, from a democratic accountability point of view, to structure elections in a way that aligned the interests of the electorate and control of government for a specific term of office. It is noteworthy in the chart above from Pew that the pattern of late has been to give presidents unified government at first but to swiftly then take it away. Given that policy is hard and takes time, that makes little sense if one wants efficacious government (however one might define that).
At a minimum, we need to stop talking like every election is about open competition over sets of ideas. We have to recognize that the outcomes of most races are essentially pre-determined and that only a handful are competitive. We tell ourselves a fictional story that elections are reflections of how well a given party has governed, when, in fact, they don’t tell us that at all–certainly not as as the democratic feedback loop we like to pretend is the case. This is exacerbated by the short amount of time we give presidents and congresses to achieve anything.
Indeed, the lack of significant democratic feedback and, hence, democratic accountability is a major reason why I am a critic of our current institutional configuration.
