The Outsider Perspective Issue 486 | The Beltway Outsiders

If you’d like to read this issue on my website, click here! If you’d like to sign-up and receive this in your inbox each week, click here! Read past issues here.

Good Friday Morning! A woman on a Caribbean Airlines flight from Kingston to New York went into labor on April 4 and delivered a baby somewhere over the Atlantic before landing at JFK. Staff radioed air traffic control for an early landing slot: “We’ve got a pregnant passenger who is going into labor at this time.” ATC’s response: “Is it out yet?” Confirmed. “Tell her she’s got to name it Kennedy.”

The fun part is the citizenship question. An immigration lawyer pointed out that the baby’s nationality depends on where the plane was at the exact moment of birth — Jamaican airspace, international waters, U.S. airspace — and that in rare cases, a baby born mid-flight can end up stateless. The U.S. has “right of soil” laws, meaning if the kid was born in American airspace, it’s American. If not, it depends on the mother’s nationality, the airline’s country of registration, and which treaty applies. Nobody seems to know the answer yet. A baby with no country and a flight path for a birth certificate.

Your Friday thought experiment: if the plane was directly over the Bermuda Triangle at the moment of delivery, does the kid belong to Bermuda, Britain, or the void?

Speaking of contested territory, this week I’m writing about the 2026 midterms. The numbers, the history, the map – links to follow.

Quick Hits:

Trump nukes the Podcastistan right. Someone joked this week that Trump ran out of things to bomb in Iran, so he put a ceasefire there and decided to nuke Tucker Carlson, Megyn Kelly, Candace Owens, and Alex Jones all in one blast. That is basically what happened. Trump issued a nearly 500-word Truth Social rant on Thursday ripping into his most influential conservative media allies over their opposition to Operation Epic Fury. Called all four of them “low IQ…stupid people” and “NUTS JOBS, TROUBLEMAKERS.” Mocked Candace Owens for the Brigitte Macron conspiracy — “Actually, to me, the First Lady of France is a far more beautiful woman than Candace, in fact, it’s not even close!” Hit Jones for Sandy Hook, said Carlson is a “broken man” who never finished college, and claimed he could get all four back on his side anytime he wanted but doesn’t return their calls because he’s “too busy on World and Country Affairs.” The four of them have been arguing that Trump was misled into launching the war under Israeli pressure and that the president abandoned his America First campaign promises. Here’s the thing: they’re all nuts, and he’s right. The podcast right has spent three years building audiences on conspiracy thinking, and the moment a Republican president does something they didn’t pre-approve from behind a microphone, they melt down. Tucker is peddling the same isolationist revisionism he’s been running since 2022. Candace is litigating whether the First Lady of France is secretly a man. Jones lost $1.5 billion for claiming Sandy Hook was a hoax. Kelly has been fishing for a lane since Fox fired her. These are not serious foreign policy critics. They are entertainers who got mad that the show went off-script. Trump calling them stupid is the most honest thing he’s said in a month.
If you don’t already, make sure you subscribe to our daily newsletters. The American Almanac delivers the day’s biggest stories every morning. Capital Digest covers the business and finance stories that affect your wallet. Conservative Legal News tracks the courts, the cases, and the legal fights that matter. And the Conservative Institute has daily commentary and analysis. All free.

Where you can find me this week

Please subscribe, rate, and review The Horse Race on YouTube — the reviews help listeners, and readers like you find me. Make sure to sign up for the Conservative Institute’s daily newsletter and The American Almanac.

Drill, Baby, Drill — But for Fertilizer – Conservative Institute

Iran Put a Bounty on a Colonel. We Got Him Back Anyway. – Conservative Institute

Forget the Ceasefire Drama. The Oil Receipts Say America Is Winning the Iran War. – Conservative Institute

The Arithmetic of November

What the 2026 baseline already tells us, what history says about a setup like this, and what still has to resolve between now and November.

Tuesday night in Wisconsin, Chris Taylor beat Maria Lazar by twenty points — 60 to 40, 905,157 to 600,044, the largest margin in a state Supreme Court race there since 2000. The AP called it at 8:36 p.m. The liberal majority on the court expands to 5-2 and holds through 2030. By Wednesday morning every cable panel was running the 2017-comparison script, treating the result as a vibe check on the midterms.

It isn’t. It’s a mood ring.

Specials don’t predict midterms; as a predictive power, they’re borderline useless. Baselines do point to real results. And the 2026 baseline — the actual load-bearing numbers, the ones voters haven’t even started paying attention to yet — is already ugly for the party in power, already consistent with a specific range of historical outcomes. Those questions are not whether Republicans take losses in November. They are how big on the House side, and whether the Senate map, which punishes both parties in different directions, produces a clean outcome or a coin flip.

Republicans are, predictably behind. Not because of what special elections signal, but because the hard numbers point that direction. So start with the baselines. What we know. What history says about them. What still has to resolve.

Start with the numbers

Generic congressional ballot: Democrats +5.8. The RCP average has Democrats at 47.5 percent, Republicans at 41.7, a lead that has ticked up over the last several weeks after holding in a tighter band for most of the winter. Out-party leads on the generic ballot are the expected setup in a midterm cycle. The party out of power almost always leads. The question is how the size of the lead compares to historical out-party leads at the same point on the calendar.

By that standard, a D+5 or D+6 April read is middle of the pack, not the top of it. In April 2018, Democrats were already running +7 to +8 and climbing toward a 41-seat wave. In spring 2006, Democrats ran +11 to +15 stretches and won 31 seats. In 1994, Republicans pulled +5 to +7 out of a standing start and blew through to +8 by the fall on the way to a 54-seat wave. In 2010, Republicans sat at +3 to +5 in the spring and still won 63 seats because approval was collapsing and the economy was bad.

A six-point April lead for the out-party is a normal midterm setup. It is also possibly a capped lead. This is a historically small lead for the party out of power in a year when the in-party’s other fundamentals are pointing this direction. My current working theory: the Democratic brand is still underwater from 2024, and that is holding their ceiling down even in a favorable environment. Americans dislike Democrats. Even Democrats dislike Democrats. CNN’s Harry Enten ran the numbers on this last week and called them “atrociously awful — a double-A for the Democrats here.” Seventy-four percent of voters overall say congressional Democrats have the wrong priorities, and Democratic voters themselves are not far behind. Net approval of Democratic congressional leaders has collapsed from +26 two decades ago to −4 today. That is the ceiling problem in one data point. The ceiling opens up fast if Democrats get their act together between now and Labor Day. It stays capped if they don’t. What we don’t know yet is whether this recent uptick to D+5.8 is the start of a real breakout or the top of the range.

Popularity isn’t needed to win midterm elections. Republicans were broadly disliked in 2010 and 2014, and still won decisive victories in Congress. However, that unpopularity can cost you in critical areas, like the Senate (the GOP left several seats on the table in 2010/14).

Presidential approval: 40.9 / 56.8, net −15.9. The RCP average has Trump’s approval at the low of his entire second term and the disapproval line at the high. The direction over the last month is down, and the slide is accelerating.

An in-party president running at net −10 to −15 heading into a midterm has historically cost his party 25 to 35 House seats. Reagan went into 1982 at roughly that territory and lost 26 seats. Bush went into 2006 at around 38 percent Gallup approval and his party lost 31 seats. Clinton went into 1994 underwater and Democrats lost 54. Obama went into 2010 at net zero and Democrats lost 63, because the economy was in the ditch on top of it. At −16, Trump is sitting at or below the floor of the historical penalty range, and the trend is still the wrong way.

The immediate catalyst for the slide is Iran. The oil and gas spike that followed the launch of the Iran operation is feeding directly into the kitchen-table discontent the rest of the dashboard was already flashing — gas prices have surged above $4 a gallon for the first time in more than three years, up more than a dollar in a month. CNN/SSRS has Trump at 31 percent on handling the economy and 25 percent on cost of living, the issue that got him elected in the first place. Reuters/Ipsos has his overall approval at a fresh 36 percent low, the worst of his second term.

Whether this is a blip or a fatal hit is the real open question on the approval line. Trump’s first-term approval ran a remarkably narrow band in the low-to-mid 40s, and his coalition refused to walk away from him no matter what the news cycle threw at them. The current number sits below the bottom of that band. If it holds here, this is roughly where he stops falling, and Republicans take the kind of midterm Reagan took in 1982. If it breaks, the comparables start looking like Bush 2006, and the seat math gets worse from there.

And there is a Biden analogue worth remembering. Biden’s approval never recovered from Afghanistan. The withdrawal was not the worst thing that happened in his presidency, but it was the moment the bottom fell out and the number stopped coming back. Trump’s Iran moment could be his Afghanistan. It could also be a four-week dip that rebounds once oil resettles and the military operation produces a cleaner story. We will know in six weeks. That is the single most consequential unknown on the 2026 board.

U3 unemployment: 4.3 percent. Per the BLS employment report, the headline unemployment rate is not a crisis number and not a rescue number. It is the number that exists while the other numbers move.

Real GDP growth: 1.4 percent annualized, Q4 2025. The BEA advance estimate clocked fourth-quarter 2025 growth at an annual rate of 1.4 percent, with full-year 2025 growth at 2.2 percent. BEA flagged that the partial federal shutdown from October 1 through November 12 subtracted roughly a full percentage point from the Q4 print. That’s a sluggish number with an asterisk — not a recession, but not the kind of growth that cancels political gravity.

CPI inflation: 2.4 percent year-over-year. Per the BLS CPI release, headline inflation is running at 2.42 percent. This is the one that matters politically. Voters do not react to core CPI, trimmed mean, or supercore. They react to the receipt at checkout and the number on the pump. Tariff pass-through is showing up in the category-level data — the political problem sitting underneath the headline. I wrote about the supply-side version of this earlier this week. The political version is simpler: the party that runs on prices and does not deliver on prices loses the midterm.

Put the five together. A generic ballot running D+5.8. An approval number sixteen points underwater and still sliding. Unemployment that doesn’t help. GDP growth that doesn’t help. Inflation voters can feel, with gas prices doing the feeling for them. Wrong-track running in the mid-60s sitting on top of all of it. Nothing on that list rescues the in-party.

What history says about a setup like this

The midterm rule is the strongest fundamental in American politics. In 19 of the last 21 midterm elections, the president’s party has lost House seats, with an average loss of roughly 25. That is a law of political gravity with two exceptions — 1998 and 2002 — both of which ran on idiosyncratic national stories: Clinton impeachment backlash, post-9/11 rally. No such story exists in 2026.

The specific historical analogues that match this baseline — approval, economic mood, generic ballot size — are Bush 2006 and Reagan 1982. They are not 1994. They are not 2018. 1994 had a mobilizing issue that ran through every district: Hillarycare plus the crime bill plus the NAFTA fight. 2018 had the Trump-resistance wave and a galvanizing story in every suburban district in the country. This cycle does not have a single wedge running through every district (for now). It has grinding, diffuse, kitchen-table discontent — same as 2006, same as 1982. That is a medium-wave profile, not a tsunami profile.

Midterms are base elections. The voters who show up in a non-presidential year are high-propensity partisans who vote in every cycle regardless of what’s on the ballot. The atypical Trump 2024 coalition — low-propensity, anti-inflation, anti-incumbent voters who broke for him in November — does not turn out in a midterm at anything like presidential-year rates. That is not unique to Trump. It is a structural feature of American elections. Obama saw it in 2010. Trump saw it in 2018. Biden saw it in 2022. The headliner carries the low-propensity coalition. The midterm strips it away.

None of which is to say Democrats are in good shape. They are not. Their brand is still underwater, the working-class Hispanic and Black drift toward Republicans is the subject of an entire running series at The Liberal Patriot, and their generic favorability is in the tank. Ruy Teixeira has been running the numbers on this for years: since 2012, nonwhite working-class voters have shifted 18 margin points away from Democrats, and the Democratic brand, in Teixeira’s read, is still defined in working-class eyes by its loudest progressive voices rather than by its median elected official. That does not save Republicans. It caps the size of the Democratic wave. A 2006-sized correction. Not a 1994-sized one.

The House math

The House math is the cleanest part of the picture. The Republican majority is five seats. Four flips and the gavel is gone. The historical base rate for an in-party loss under these conditions is five to six times that.

Put a box around the central case: a 20-to-30 seat loss. Bush 2006 dropped 31. Reagan 1982 dropped 26. Carter 1978 dropped 15 in a softer environment than this one. Twenty to thirty is where the fundamentals point, and the fundamentals would have to move visibly before that range moves.

What moves it smaller: inflation rolls back hard between now and August, approval recovers eight to ten points by Labor Day, the economy quietly accelerates, no foreign-policy shock. What moves it bigger: a recession hits before summer, approval breaks through the low-40s floor, Iran deteriorates past the current holding pattern, or tariff pass-through shows up harder in the CPI prints. The upside and downside are not symmetric — the margin of error runs wider to the downside than the upside, because the floor-breaks scenario is worse than the floor-holds scenario is good.

What we don’t know is which side of that 20-to-30 range we land on. That is a late-summer call, not an April one. The default assumption has to be substantial losses. The House is the likelier casualty than the Senate, and anyone giving you a cable-friendly seat number on the nose in April is selling, not analyzing.

The Senate is its own map

Senate control is a different conversation because the seats in any given cycle are fixed six years in advance. The map drives half the outcome before the environment touches it. The 2026 map is ugly for both parties in different ways.

The top-line number that matters: eleven incumbent senators are not seeking re-election — four Democrats and seven Republicans — per Ballotpedia’s March 24 tally, the most in a single cycle since 2012. Open seats in a bad environment are harder to hold than incumbent seats, and the GOP is carrying more of them.

The races that decide the chamber (ratings throughout this section from Cook Political Report, cross-checked against Sabato’s Crystal Ball and Inside Elections):

North Carolina. Thom Tillis is retiring. The state is R+1. Cook moved the race to Toss Up the day Tillis announced, and Sabato and Inside Elections followed. It is the top Democratic pickup target on the board.

Maine. Susan Collins is running. The state is D+4. Every rater has it Toss Up or Tilt R. Collins has survived worse environments than this one, but every cycle gets harder, and she is carrying a national climate she did not pick. The other thing worth knowing about Maine: the state has notoriously bad polling. Whatever the public numbers say here, discount them. Collins could cruise. Collins could lose. Neither outcome would surprise anybody who has watched a cycle there.

Georgia. Jon Ossoff is defending. First-term Democrat, R+1 state. This is probably the single most important Senate race on the map — if Ossoff holds, a Republican Senate majority probably holds with him. If Ossoff falls, the chamber is genuinely up for grabs.

Michigan. Gary Peters is retiring. Open seat, Tossup to Tilt D, candidate quality on both sides not yet finalized.

New Hampshire. Jeanne Shaheen is retiring. Open seat, Tossup to Tilt D, same fields-not-set caveat.

Iowa. Joni Ernst is retiring. R+6 and normally safe, but open seats in a bad environment stop being automatic, and Democrats recruit well when the map gives them an opening.

Texas. John Cornyn and Ken Paxton are headed to a May 26 Republican runoff. GOP insiders are already openly fretting about Paxton as a general-election candidate, and with reason — he is viewed as uniquely weak in this race for a list of obvious reasons. The Democrat is state Rep. James Talarico, who beat Jasmine Crockett in the primary on a “progressive-Christian pitch” — a religious Beto O’Rourke, merging progressive policy with Christian theology, which is the exact persona a radical progressive socialist cooks up to run in Texas, thinking, “Yes, this is what the right leaning people want.” Should be safe GOP. Probably still will be. But if Republicans nominate the unelectable version of the Cornyn-Paxton choice, Texas becomes a seat Republicans could lose without Texas having turned blue.

Four to six of these contests hinge on candidates who haven’t declared or haven’t won their primaries. Anyone giving you a Senate seat count in April is guessing. The honest read is that the chamber probably lands in a 50-50 or 51-49 range in either direction, and calling which direction today is not defensible. Candidate quality in North Carolina, Michigan, and New Hampshire, and the outcome of the Cornyn-Paxton primary in Texas, are the four known unknowns that determine where it resolves.

The gerrymandering objection

Both sides are going to play this card between now and November. Democrats will argue that the House is rigged by Republican state legislatures and the fundamentals don’t matter because the maps are cooked. Republicans, if the wave shows up, will argue the same thing about blue-state maps in Illinois, Maryland, New York, and California. Neither side is going to be entirely wrong, and neither side is going to be mostly right. And the argument itself is not new. It has been the standard losing-side explanation in every cycle for the last fifty years.

Both parties actively try to rig their maps. That is true. It has always been true. It was true in 1962. It is true now. The only thing that changes cycle to cycle is which side has the bigger advantage at any given moment.

The current Republican structural advantage is real but modest — two to three seats’ worth of baked-in tilt, and it has been priced into the fundamentals models for years. It is not going to override a generic ballot running six points underwater plus an approval number sixteen points underwater plus gas above $4. The map is big. The math is bigger.

Gerrymandering matters at the margin. It has never overridden the fundamentals in a midterm like this one. It is not going to start now, regardless of which side is making the argument.

What to watch

Six things between now and Labor Day will tell us where the range lands.

First, the Trump approval floor. Does the low-40s support level hold or break? Does the Iran-driven slide rebound or calcify? The next four to six weeks of averages are the read, and the 36-to-37 percent floor on the low-end single polls is the number to watch — if it shows up on the RCP average, this is Biden-Afghanistan, not a four-week dip.

Second, the generic ballot. Does D+5.8 expand toward D+7 or D+8 — a 2018-profile wave — or stay pinned in the D+4 to D+6 range that looks like a 2006 profile?

Third, CPI. Do higher prices show up in the spring prints, or stay muted through the summer?

Fourth, the Q2 2026 GDP print. The historical fundamentals models weight election-year first-half growth heavily, and the Q2 release is the last major economic data point before voters start paying attention.

Fifth, Senate candidate fields. North Carolina Democratic field, Michigan Democratic field, New Hampshire Democratic field, Texas Republican primary. These are the races that cannot be called until the candidates are real.

Sixth, the wrong-track average. Does the mid-60s reading come down as the economy clears, or stay elevated as inflation bites? Wrong-track is a leading indicator the rest of the dashboard catches up to.

It’s still the economy

The map hasn’t changed in eighty years. Either the party in power delivers on prices, or the arithmetic plays out the way it has played out in every comparable midterm since 1946.

We’ll run this exact baseline again in July, and again in September. These numbers the true signal, while all the special elections and one-off polls are noise.