Stupid predictions N0.2: The Senate flips back to Republican

I love to use math and history to predict the future, no matter how foolhardy. Earlier I predicted a Republican takeover of the US House of Representatives, with that chamber going from a 10-seat (of 435 total seats) Democratic advantage to a 50-seat Republican advantage. While I called that a 60-seat swing, that was in the margin and not the seats. Someone who uses math a lot should know that each lost seat is a won seat, so it only takes a 30-seat change for Republicans to end up with a 50-seat advantage.

Now let’s take a look at the Senate. Currently there are 48 Democrats and 50 Republicans. However, two independents join the Democrats in votes and Vice President Kamala holds a tie-breaking vote that gives Democrats very narrow control.

Reporters and politicians believe a party can accomplish a lot when they control the presidency and both houses of Congress, though some research suggests otherwise. Once we conclude that Republicans will control both houses of Congress, called “divided government,” we’ll speculate on how that will change national politics.

Senate races are different, but often similar

The chart below shows shows the percentage of seats held by the President’s party after every election back to 1898. Senate and House races generally follow the same pattern, but results are not identical. The House and Senate moved the same direction in 43 of these 61 elections (70 percent), and in 10 elections in a row up until the last two. (Note that I’m ignoring members who aren’t Republican or Democratic, since I can’t tell in some years how the independents and third-party candidates lined up).

In spite of the general similarity in results, there are a number of important differences between House and Senate races, each of which makes election dynamics different in the two houses.

• While all 435 seats in the House of Representatives are up for election every two years, only a third of Senate seats are, because Senators serve six-year terms. This year that means 35 Senate elections, since there is a special election in Oklahoma.

• Senate races often react less to short-term political issues. That’s because the Senators running for re-election were last elected in 2016, when Donald Trump was elected president, while House members running for re-election were elected in 2020, when Biden was elected. That may not matter a lot since neither of these presidents had the normal “coattail effect” that increased their party’s numbers in Congress.

• Senate races are usually closer, since they are elected statewide. Most house districts are designed to heavily favor one party or the other, while states as a whole are more balanced politically, so either party has a decent chance at winning a Senate seat in any state.

• This year, a couple factors favor the Democrats in the Senate, but not the House. First, five of the six retiring senators are Republicans, while 22 of the 32 retiring representatives are Democrats. Further, just 14 of the 35 Senate seats up for election this time are held by Democrats, so Republicans will have to work very hard just to stay even. Since every House seat is up for election and since Democrats have a very small margin, only a few more Democratic than Republican seats are up for grabs.

• The Senate is only Democratic-controlled now due to a fluke. Democrats grabbed both Georgia Senate seats in 2020, and thus control of the Senate, due to that state's weird runoff election held when President Donald Trump’s election denials had reached a new level of absurdity, at least to that point.

• Of course the structural elements favoring Democrats could be canceled out by continuing poor approval ratings for Democratic President Joe Biden and the lowest public confidence in the economy since the Great Recession.

What are the experts saying?

Given all these factors, it’s not surprising that those who get paid to watch politics (my resume is still in your spam folder, guys!), can’t agree on how the Senate races will turn out. Real Clear Politics continues to predict a slight Republican win–a gain of two seats, Nevada and Georgia–in the Senate.

FiveThirtyEight, on the other hand, gives Democrats a 66 percent chance of keeping their very narrow control of the Senate. They also suggest Republicans have a good chance in Nevada and Georgia, but give Democrats the nod in Pennsylvania and a good shot in Wisconsin.

The Cook Political Report is right in the middle. They show the Georgia and Nevada seats now held by Democrats as tossups and the Pennsylvania and Wisconsin seats now held by Republicans in the same category. Cook doesn’t predict winners, at least in front of their paywall. So their range of likely outcomes is between a two-seat gain for either party.

One thing all these experts have in common is they are seeing a slight reverse in apparent Democratic gains that resulted from the end of Roe v. Wade and some real Democratic accomplishments over the summer. The generic congressional ballot shows the same pattern. I don’t see anything on the horizon for the Democrats to make further gains, and continued worries about inflation and growing concern that we are heading for a recession won’t help the party in power.

Look for a small but important change in the Senate

Democrats’ narrow margin in the Senate is in trouble. Unlike the House races, though, it’s hard to establish that with data (other than polling, which may undercount the Trump wing (which is the only wing at this point) of the Republican party. I looked at a number of indicators going back a few years to a few decades and didn’t find any meaningful relationship between most of them and Senate election outcomes. Presidential approval and the results of last year’s Virginia elections explain two percent or less of Senate results.

The generic congressional ballot question (would you vote today for a Democrat or Republican for the House of Representatives?) is a good indicator, explaining about half of the variation in Senate seats. Based on the June generic ballot, the 2.5 point lead for Republicans would predict a Republican gain of 5 seats, changing the Senate from 50-50 (with a Democratic tiebreaker) to 55-45 Republican. That’s quite a bit more than the experts are saying, and in looking at individual states I am not sure how they'd do it. But it's certainly not impossible.

We can also use the historical relationship between House and Senate changes to predict the upcoming election. As noted earlier, Senate and House seats have moved in the same direction in 70 percent of the 61 elections since 1900. They’ve moved together a bit more recently (eight of the last ten elections) but a little bit less in first midterm elections (68 percent). Over the long run, when the President’s party loses on both houses of Congress, they lose a Senate seat for every seven lost House seats. That’s roughly the ratio in midterms as well. The last ten elections have been bloodbaths, though, with the President’s party losing a Senate seat for every four lost House seats. If my prediction of a 30-seat House loss for President Biden and Democrats is correct, the overall and midterm seat loss ratio would translate to a Democratic Senate loss of 4 seats. If the recent election ratio were to hold, though, they’d lose seven seats.

As with the House, I prefer to let history be our guide. The two historical measures suggest a Democratic loss of four or five seats in the Senate. I’m going to hedge a bit based on gut feel and go with four, changing the Senate from 50-50 (with a Democratic tiebreaker) to 54-46 Republican. That’s quite a bit more than the experts are saying, and in looking at individual states I am not sure how they’ll do it. But people rarely go wrong betting against Democratic presidents, particularly in midterms. As the chart below shows, the last Democratic President to gain Senate seats in the first midterm election was John Kennedy. The three Democrats since then have taken big hits and I don’t see Biden faring much better.

The new Congress isn’t going to be fun. Here’s how it could look

Divided government is messy, and it won’t get any cleaner with our deeply divided electorate and an almost instant switch of attention to the Presidential election and whether we’ll have our choice between two 80-year-olds or something different. Here’s some things to watch for if Republicans control both houses of Congress while Democrats hold the White house.

• Hunter and Hillary and Harris and so many others. Republicans in both houses will go absolutely nuts conducting investigations. Their party doesn’t have to involve Democrats in investigations so they are a great tool to fling mud at Democrats and stay in the news. We can expect lots of special committees, televised hearings, and crazy accusations. I’m betting the House will impeach Biden, Harris, and Attorney General Merrick Garland at least once, though the Senate won’t be able to muster votes to convict anyone, so they may not even hold trials.

• Budgets will be showdowns over everything except the actual budget, and we’ll see at least a couple of government shutdowns while this Congress is in session.

• There will be some worthwhile legislation passed, though it’s hard to guess what it might be, and it will depend on whether Republicans are willing to share any credit with Biden and a few Democratic senators. Republicans will pass many laws, including abortion bans, attempts to sway elections, and a few more tax cuts for the rich, but they won’t be able to override Biden vetoes, so it will just be for the theater.

• Collins (and Murkowski and maybe Romney) are the new Manchin. The narrow Democratic lead in the Senate has made the most conservative Democrat, Joe Manchin of West Virginia, the most important vote in the Senate. While the Republicans may have a narrow lead after this November’s elections, they’ll need nearly every member to get even simple laws passed. That means their least conservative members, Susan Collins of Maine, Lisa Murkowski of Alaska, and Mitt Romney of Utah, will determine how votes come out. They’ll decide if a Biden nomination for the Supreme Court gets confirmed, and they’ll decide whether to shut down the government over budget disagreements.

So, good luck to all candidates, pundits, voters, and to all of us who will have to live with the result. If I’m right, which we’ll check back on next month.