The Moderate Is Whoever Quotes the Second Number
In September, 66% of Americans told KFF they favored Medicare for All. The question described it as “a national health plan … in which all Americans would get their insurance from a single government plan.”
In the same month, Axios-Ipsos asked about the same policy and got 43%. Their wording said “government-run health insurance to all.” The wording added four words the first question left out. Those words were “largely replacing private insurance.”
Same policy, same country, same month, and the two polls land twenty-three points apart. One question asks whether you want the thing. The other tells you what it costs and whose coverage goes away.
- KFF fielded Sept 16–21
“a national health plan … in which all Americans would get their insurance from a single government plan”
- Axios-Ipsos fielded Sept 25–28
“government-run health insurance to all, largely replacing private insurance”
Medicare for All support measured by two pollsters in the same month. KFF, which described a single government plan, found 66 percent. Axios-Ipsos, which described government-run insurance largely replacing private insurance, found 43 percent. The gap is 23 points.
What I got wrong
In April I built an index from 85 polls across eight issues, and I named six races as its test. The test was never about the people. It was about whether a platform the press calls far-left can win a general election. Two of the six were decided before the piece ran. One candidate never advanced, one dropped out, and one result never came back. That left one clean test of a candidate, which isn’t a test. It left two clean tests of a platform.
Maine is the race I would have scored a miss. Graham Platner, the presumptive nominee in April, dropped out. Troy Jackson took the nomination and ran on the same platform. Medicare for All sat at the top of it. He won in November. If the bet had been on Platner, Maine would be a loss. It was a bet on the platform, so Maine is a win. That also makes it a cleaner test than Michigan. There the candidate and the platform arrived together. In Maine the name changed and the platform still won the state.
My first read was that this was the greatest rout of Republicans in my generation. That was hyperbole, and the House count doesn’t support it. Democrats sat at D+18 to D+30 in called seats, with about D+43 projected. Republicans took 63 seats in 2010 and 54 in 1994. Democrats took 41 in 2018.
The Senate is where the claim holds. Democrats won 56 to 44, a net gain of nine and the largest swing since 1986. Alaska and Mississippi both went Democratic. They also took five governorships, their first majority of governors since 2010.
The exit polls say what moved voters. They named the economy at 39%, the state of democracy at 32%, and immigration at 14%. They called themselves 25% liberal, 41% moderate, and 34% conservative. Fifty-four percent viewed socialism unfavorably. NBC’s summary was “a blue tsunami, not a progressive earthquake.”
The war did the rest. Sixty-two percent called the Iran war the wrong decision. Opposition hardened as the war ran, from 43% in its first week to the high sixties by July. Voters met it as a price. Sixty-one of sixty-five political ads about the war were about the cost of living. The window was September to mid-October. Gas was $2.94 a gallon the week before the war started and $4.15 the week before the election. It did not rise into the vote. It had been above $4 for seven months.
The price, on a ballot
Oklahoma voted on June 16. State Question 832 would have raised the state minimum wage to $15 by 2029. It lost 44.6 to 55.4, in a low-turnout June primary with 630,488 votes cast. It carried Oklahoma and Tulsa counties and lost rural Oklahoma.
The attorney general’s office wrote the ballot title. It warned that the measure would have “a fiscal impact on the State.” Voters read that line first. Then they read the wage:
“a revenue increase by new taxes or elimination of existing services.”
That is the second number, and someone wrote it on purpose. Nationally, raising the wage barely moves anyone. 67% back $15, 64% back $17.50, and 60% back $20. The whole slide from $15 to $20 costs seven points. In Oklahoma, where the tax sentence came attached, $15 got 44.6.
One state isn’t a national result. It is a red state, a June primary, 600,000 voters. But this is one of the few places in the data where someone wrote the price down. Then the public voted on it. California wrote a different price. Proposition 40, a one-time 5% levy on net worth above $1 billion, trailed 44.4 to 55.6 after being outspent roughly 250 to 1.
National, by the hourly rate asked for
- $15 67%
- $17.50 64%
- $20 60%
Oklahoma, the same $15 on a ballot
- $15 44.6%
Yes vote, with a ballot title naming the tax attached to it
National support for a federal minimum wage falls only slightly as the rate rises: 67 percent back $15, 64 percent back $17.50, and 60 percent back $20. In Oklahoma, where $15 appeared on a ballot with a title warning of a tax increase, it received 44.6 percent.
The ledger
AIPAC and its super PAC, United Democracy Project, spent $120 million across 333 of 473 seats. The marquee races went against it. It spent $30.6 million on Michigan’s Senate primary, and El-Sayed won it. NJ-11 and CA-14 went to the candidates it opposed. Most low-profile races went its way. On October 21 it stopped endorsing candidates. An internal letter called 2026 “one of the most challenging environments in recent memory.”
On election day, Trump told Fox News the United States would end all military aid to Israel. He said the cut would land by the end of 2026. That aid runs about $4 billion a year. He gave the reason out loud. Aid had “become a Democratic talking point,” he said, for candidates like El-Sayed. A July poll had found 58% wanted that aid ended outright.
No floor vote cut that aid. It came by executive action, after the counting, and the stated cause was a slogan. Whether the polling forced the cut or let him defuse the issue can’t be known. On the one poll that asked about vote choice, a pro-aid candidate led 45 to 35.
Neither brand lost
El-Sayed won Michigan 52.5 to 45.5. His slogan was “Money out of politics, money in your pockets and Medicare for All.” It leads with money. Then it names the program the money would pay for. He won with $30.6 million spent against him. Shapiro won Pennsylvania 60 to 38 as a moderate.
Both of those are true, and the returns can’t tell us which cause carried which. What they can say is narrower and worth more: none of it was disqualifying. The index had measured the establishment moderate. On seven of eight issues, he sat to the right of the public. The public did not vote on those seven. The 43% isn’t a ceiling on what a candidate may propose. It is the price of proposing it without saying who pays.
Who says the number first
The moderate isn’t a fraud and isn’t a coward. He quotes the second number because it is the only number anyone lets him say out loud. The index was right about what people want. I was wrong about what the country would vote on. El-Sayed won by putting the money in his first clause, before anyone could put a price on it for him.
People set prices, in rooms, for reasons, and we are the ones who pay them. Quoting the second number isn’t a concession. It is the start of an argument about who sets it. That argument is open to anyone who says the number first.
