The Outsider Perspective Issue 498 | 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! America has a parasite problem, and it is coming for the salad bar. The CDC counts nearly 7,000 cases of a bug called Cyclospora across 34 states this summer, well past the old U.S. record of about 4,700 set in 2019. The parasite produces what the tabloids are delicately calling “explosive diarrhea,” and the New York Post spent the week hunting for the source: Taco Bell.

The prime suspect is not a burrito. It is lettuce. Public-health officials in Michigan have pointed to salad greens as the likely culprit, with Taco Bell pulled into the investigation because they use the main supply company involved. Somewhere a man who has eaten nothing but Doritos and Mountain Dew since 2011 is feeling deeply validated (I plead the fifth).

My takeaway, as I mentioned on Twitter: This just means my plan to eat more bacon is proving to be the safest bet of all.

The Babylon Bee, naturally, has not let this go. In the past week it has traced the epidemic to a “gain-of-function lab at the Taco Bell Institute” and reported that theologians now agree the outbreak is “an outpouring of God’s wrath on vegetarians.” I’m not sure the last one is incorrect, especially regarding vegans.

This week, I’m setting down the midsummer marker on the 2026 midterms. In April I laid out the baseline and promised to run it again now. And today I’m covering why the House and Senate are drifting in opposite directions – links to follow.

Quick Hits:

Tennessee could hand the Supreme Court the case Oklahoma couldn’t. City Journal’s Ray Domanico makes the case that a new Christian charter school outside Memphis is shaping up to be the next landmark fight over religion and public education. The last one died on a technicality. When Oklahoma’s bid for a religious charter reached the Court, Justice Barrett recused, the justices deadlocked 4-4, and the biggest church-state question of the decade was left unanswered. The issue itself is simple and unavoidable. A charter school is a public school a state chooses to authorize. If the state hands that benefit to secular operators, can it categorically slam the door on religious ones? The Court has already told states that excluding an institution from a widely available public benefit just because it is religious is unconstitutional discrimination. Tennessee looks like the vehicle that finally forces the question with a full bench on the issue of charter schools.

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.

Ask a Democrat to Name an American Killed by an Illegal Immigrant. Watch What Happens. – Conservative Institute

The Left Can’t Define a Woman. They Sued This One for Refusing to Abort a Baby. – Conservative Institute

Mexico Can Name 17 Of Its Dead. It Can’t Name Dead Americans. – Conservative Institute

Republicans Got Their Best Six Weeks of the Year. The Map Didn’t Move.

The midsummer check-in on the 2026 baseline: what moved since April, and why the House and Senate are drifting in opposite directions.

The July polls handed Republicans their best headlines of the year. RMG Research put the generic ballot at Democrats plus two this week. Harvard-Harris matched it. So did 2Way/HarrisX. Sean Trende noted that the RealClearPolitics average had fallen from Democrats plus eight to plus five in six weeks as soft Republicans came home, and the party’s operatives spent the week telling anyone who would listen that the wave is dead.

Six good weeks in July is a mood ring, not a verdict. I set the 2026 baseline in April and promised to run it again at midsummer and again in September, watching which numbers moved and which held. This is the midsummer read, and it points two directions at once.

The numbers moved at the margins and held where it counts. The House still leans Democratic, and the lead is shrinking. The Senate is sliding toward Republicans. The reason is the same in both chambers: the fundamentals are still grinding against the party in power, and the Democrats cannot stop nominating candidates who lose races they were already winning. Republicans think the tightening polls are a reprieve. Democrats think the environment is a gift. Both are wrong.

Republicans clawed back six weeks. They did not claw back the map.

Generic congressional ballot: Democrats +5.0, essentially where it sat in April. The RealClearPolitics average has Democrats at 48.1 and Republicans at 43.1. In April the same average read Democrats +5.8. In between, it ran higher over the early summer while Trump’s approval bottomed out, then compressed as Republicans consolidated. Nate Silver’s Silver Bulletin aggregate runs a point friendlier to Democrats at plus six. The Democratic lead is no larger than it was three months ago. It briefly ran higher, and Republicans have pulled it back to the spring level. That is a real recovery. It is not a realignment.

The July spread underneath the average runs wide, from Democrats plus two to Democrats plus nine, which is why the Republicans quoting the plus-two polls and the Democrats quoting the plus-nine polls are both cherry-picking. The average is the number. And a five-point out-party lead in midsummer is small by the calendar. In April 2018 Democrats were already running plus seven and climbing to a 41-seat wave. In spring 2006 they ran plus eleven to plus fifteen. Republicans pulled plus five to plus seven in 1994 and blew through to a 54-seat wave; they sat at plus three to plus five in the spring of 2010 and won 63 seats because the president was underwater and the economy was bad. A capped lead in a favorable year means one thing, and it is the same thing it meant in April: the Democratic brand is still holding the party’s own ceiling down.

Presidential approval: 41.0 approve, 56.4 disapprove, net −15.4. The RealClearPolitics average has Trump almost exactly where he sat in April, when the Iran operation and four-dollar gas were dragging him toward the floor. He found the floor. Approval bottomed near 39 in the spring and has ticked up to 41, the best stretch he has had since then. Silver’s Silver Bulletin average has the net a point and a half worse and climbing off the low. The low-40s band that Trump has held for a decade held again. That was the open question in April, and the answer so far is that the coalition did not walk.

It does not rescue him. A president sitting at net minus fifteen going into a midterm is in the penalty zone. Gallup finds that presidents above 50 percent approval lose an average of fourteen House seats, and presidents below it lose an average of thirty-six. Reagan went into 1982 near this number and lost twenty-six seats. Bush went into 2006 below it and lost thirty. A recovery off the floor keeps Trump out of a 2006 collapse. It does not lift him out of the penalty zone.

U3 unemployment: 4.2 percent. The June employment report held the headline rate steady on a soft print of 57,000 jobs. It is not a crisis and not a rescue, and it held flat while the numbers that move votes moved.

Real GDP growth: 2.1 percent, first quarter 2026. The BEA third estimate revised the quarter up from 1.6 percent, a reacceleration off a fourth quarter that grew only half a percent. It is a better number than the one I cited in April. But it is not the kind of growth that cancels political gravity, and the second-quarter print that the fundamentals models weight most heavily is not out yet.

CPI inflation: 3.5 percent year over year. This is the number that moved, and it moved the wrong way. The June CPI report put inflation at 3.5 percent, up from 2.4 in April and well above the Federal Reserve’s 2 percent target. Core inflation is 2.6 percent and sticky, and Morningstar reports that economists expect tariff costs to push it higher into the second half of the year. Voters do not grade inflation against a target. They grade it against the total at the grocery checkout, and that total keeps climbing. Prices are warmer than they should be and warmer than they were this spring, and that is the one input on this board that decides midterms.

Put the five together. A generic ballot back to its April level. An approval number that stopped falling and stayed sixteen points underwater. Unemployment and growth that do nothing to help. Inflation that voters can feel, a point hotter than spring. Nothing on that list moved the party in power out of a midterm.

History does not take the summer off.

The midterm rule is the strongest fundamental in American politics. The American Presidency Project’s data shows the president’s party has lost House seats in twenty of the last twenty-three midterms, an average loss near twenty-eight, and the only exceptions ran on idiosyncratic national stories. There is no such story in 2026.

The environment looks like 2006 and 1982 in kind, not 1994 or 2018. 1994 had Hillarycare and the crime bill running through every district. 2018 had the Trump-resistance wave in every suburb in the country. This cycle has no single wedge, just grinding, kitchen-table discontent with inflation ticking back up underneath it. It is the same diffuse dissatisfaction that ground Republicans down in 2006. That was not a mild year. Democrats gained thirty-one House seats and took both chambers; Republicans lost twenty-six seats in 1982. What is different now is the size of the Democratic lead. In the spring of 2006 Democrats ran eleven to fifteen points ahead on the generic ballot. Today they lead by five, because their own brand is holding the number down. But a smaller lead is not a smaller danger. In 2010 the out-party led by three to five points in the spring and still won sixty-three seats, because the president’s approval collapsed and the economy went with it. The lead is capped. The downside is not.

The comforting story Republicans are telling themselves about the July numbers runs into the calendar. The public is barely paying attention right now, and the people who are not paying attention break against the party in power once they start. Political scientist Justin Algara told the Niskanen Center that the two parties sit near parity while the public is still “squishy,” and that as the election nears “the bottom really starts to fall out” against whoever holds power. Labor Day is the start of that sprint, not the end of it. The tightening Republicans are celebrating in July is the part of the cycle that historically gives way, not the part that holds. Democrats’ actual hope is narrower and more specific: that independents, who have drifted since 2024, swing back to them in the fall the way out-party voters have in past cycles. That is a real possibility. It is also the thing that has not happened yet.

The House still leans Democratic, and the lead keeps shrinking.

The Republican majority is thin enough that Democrats need a net of three seats to take it. Every rater leaves control to the toss-ups. Cook Political Report counts eighteen of them, with 212 seats leaning Republican and 205 leaning Democratic, so neither party reaches 218 without them. The Hill reported that Sabato’s Crystal Ball has Democrats narrowly favored. Alan Abramowitz’s generic-ballot model has called House midterms within ten seats for decades, and this year it projects a Democratic gain around twenty-three seats at a six-point ballot, with Republicans needing to be tied or leading to have a real shot at holding.

The models favor Democrats. The rated map is a genuine toss-up with a slight Republican edge. The gap between those two things is mostly the maps, which is the objection every panel is about to run.

Gerrymandering did not repeal the fundamentals.

The mid-decade redistricting war was real, and it netted out smaller than either side claims. Cook Political Report estimates Republicans gained about five net seats from the whole scramble. Texas drew five new Republican seats and the Supreme Court upheld the map in April. Florida added four. But California’s Proposition 50 answered with five Democratic seats, and the Democratic counter-maps that were supposed to even the rest of it out collapsed. Virginia’s amendment was struck down in state court and the Supreme Court refused to revive it; Maryland’s died in a Democratic state Senate. Five net Republican seats against a three-seat majority is a real cushion, and it is the reason the rated map is closer than the models.

It is a cushion, not a shield. Five seats do not override an approval number sixteen points underwater, a generic ballot Democrats lead, and inflation moving the wrong way. Gerrymandering matters at the margin. It has never overridden the fundamentals in a midterm like this one, and a redistricting edge that Abramowitz’s own model says shaves only about five seats off the Democratic number is not going to start now. The House still leans Democratic. The maps just made the lean narrower and the majority, whichever way it falls, thin.

Democrats blew up their best Senate race themselves.

The Senate is a different map with a different logic, because the seats are fixed six years in advance and the map drives half the outcome before the environment touches it. Republicans hold it 53 to 47, and Democrats need a net of four to take it. On paper the exposure favors Democrats: eleven senators are not seeking reelection, seven of them Republicans, and the deaths and primary losses this cycle have fallen on the Republican side too. None of it is helping them, because the Democratic side keeps detonating.

Maine was the easiest Democratic pickup on the board, and Democrats blew it up. I have written enough about Graham Platner. The story that matters now is what replaced him, which is nothing better. His campaign collapsed in July, and Maine Democrats meet at a convention in Bangor on July 25 to pick a replacement from a field of more than a dozen candidates. The early polling has the leading replacements running even with Susan Collins at best, where Platner had actually led her. Democrats took their clean shot at Collins in a favorable year, wasted it on a candidate they had to disown, and are restarting in late July with a weaker hand and no nominee yet. That is not bad luck. That is a party that cannot vet its own front-runner in the one race it could least afford to lose.

Michigan is the next grenade. The open seat there runs through an August primary between Abdul El-Sayed, the Sanders-backed progressive, and Haley Stevens, the centrist that Cook Political Report calls the stronger general-election candidate because she is the one who can hold the seat in November. If the Democratic primary electorate picks the candidate who thrills the base over the candidate who wins the state, it will have done to Michigan what it just did to Maine.

North Carolina is the one pickup Democrats can count on. Roy Cooper leads Michael Whatley by seven for an open seat Republicans are vacating, and that race is theirs to lose. The rest of their offense is a set of mirages. Sherrod Brown runs close in the early Ohio polling, but Brown already lost this state in 2024, and a candidate the voters just fired, running in a state Trump has carried three times, is not a pickup waiting to happen. Ohio tilts back to appointed Senator Jon Husted. Iowa’s open seat sits in a state that votes reliably red now, and the tie in today’s average resolves the way red-state open seats resolve. Texas is the perennial Democratic fantasy. Ken Paxton is a weak nominee who runs even for the moment, but Texas does not send Democrats to the Senate, and it will not start in 2026. Strip out the mirages and Democrats are left holding Georgia, where Jon Ossoff leads Mike Collins by under three, and surviving Michigan, a coin flip they could hand away in their own August primary.

That is the whole problem. To win the Senate, Democrats have to hold Georgia, survive Michigan, win North Carolina, and then run the table on a set of red-state races that are all tilting the other way. Near-perfect runs do not happen, and Democrats are not a near-perfect party right now. None of this makes Republicans a juggernaut. Paxton is a mess, and Sabato moved three races toward Democrats in June. But close races in red states break red, and a party that keeps tripping over its own candidates does not run the table. The Senate is slipping to the Republicans, and the lean is hardening.

Five questions will settle where this lands.

First, the primaries are still running. Maine picks a nominee at its July convention, Michigan votes in August, Minnesota and Alaska follow. Half the Senate map is not set, and no one should be putting seat numbers on the board until it is.

Second, Trump’s approval floor. It held near 41 through the summer. Whether it holds there through the fall, or slips again as tariff prices bite, is the difference between a 2006 result and something milder.

Third, inflation. The June headline fell on cheap energy, but the year-over-year number is climbing and the tariff pass-through is still coming. If the summer prints keep rising, the one input that decides midterms keeps cutting against Republicans.

Fourth, whether independents come home to Democrats in the fall. That is the swing that turns a capped five-point lead into a real wave, and it is the Democrats’ whole theory of the case. It has not shown up yet.

Fifth, an event with no historical comparison at all. Trump announced that Republicans will hold the first midterm convention in the party’s history, in Dallas in September, built around a “Great American Comeback” theme. No party has ever staged a national convention in a midterm year, so there is no base rate for what it does. It could nationalize the election and rally the low-propensity Trump coalition that normally stays home in a midterm, which is the entire point of holding it. It could also hand the out party a single, focused target at the exact moment the public starts paying attention. Nobody knows, because nobody has tried it.

It is still the economy.

The map has not changed in eighty years. The party in power answers for the economy, and either it delivers on prices or the arithmetic plays out the way it has in every comparable midterm since the war. Republicans are not delivering. Inflation is warmer than it should be and the president is still underwater. A good six weeks of July polling changes neither. That is why the House is probably gone.

But a bad year for Republicans is not a good year for Democrats, and that gap is the story of 2026. Democrats will likely win the House because the fundamentals hand it to them; gravity does the work, not the party. They are losing the Senate because no environment, however favorable, fixes a party that nominates a man it has to disown in the one state it had to win. One chamber is a verdict on Trump’s economy. The other is a verdict on the Democrats themselves.