Most people don’t believe this, but if forced to declare my own personal beliefs, I do:

The Republican Party stole the 2004 election for George W. Bush.

There, I said it.

I’m not talking about their usual dirty tricks and voter suppression tactics — which, yes, they did. I’m talking about changing the results, electronically or otherwise. The evidence is in the exit polling.

Nobody talks much about exit polling anymore, but that’s because of 2004. And here’s why: In a briefing to subscribers on the afternoon of Nov. 2, 2004, the company conducting the national exit polls told its clients that Democratic Sen. John Kerry was leading in the key battleground states and appeared to be en route to winning more than 300 electoral votes.

Yet in the early morning hours of Nov. 3, Kerry conceded after a close loss in Ohio.

The exit polls had been wrong, the media said, and then offered this ad hoc explanation: Kerry voters were more likely than Bush voters to reveal their choice to pollsters.

Utterly unproven. No data to suggest it. Yet it’s universally accepted by the American Establishment.

Some weeks after the election, someone gave me a statistical analysis of the gap between Ohio’s exit polls and its official count. The exit polls showed Kerry winning with a margin above 4 percent. Ohio’s Secretary of State declared Bush the winner with a margin of 2.5 percent.

That isn’t a huge swing in generic opinion polling. But in polls of people walking out of their polling place? It’s practically unheard of.

I don’t have that analysis anymore, but here’s what caught my eye: The mathematics professor examining that deviation concluded that the odds against that shift weren’t measured in hundreds to one, but in thousands to one.

So I called a math professor at the local college. Told him I wasn’t qualified to critique the analysis and asked him if he’d take a look. The professor — with whom I was friendly, thanks to my science coverage — said sure. He seemed legitimately interested, and promised to give me a call when he was done.

That call never came. Neither would he return my emails nor repond to my voice messages. But I did eventually run into the guy: I was standing in a busy, narrow hallway at the College of Charleston, talking to a source for another story, and I looked up to see that mathematics prof headed my way.

When I raised my hand in greeting, all color drained from his horrified face. And when I took a step towards him, he turned and fled. And by fled, I mean he ran. Imagine the scene: A hallway crowded with students, and a slightly overweight, white-bearded math professor is hustling through it like the Hounds of Hell are on his trail.

Make of that what you will, but here’s what I believe: There are some facts that are so far beyond what we want to believe about reality that we simply refuse to account for them. It’s career suicide — not just for tenured math professors, but for journalists, too.

Then again, I don’t have a career anymore, so here goes:

Professor Steven F. Freeman, who later wrote a book on the topic, calculated that the odds against the exit polling in Ohio, Pennsylvania and Florida being so far off the official count were — drum roll, please — one in 660,000.

”As much as we can say in sound science that something is impossible,’it is impossible that the discrepancies between predicted and actual vote count in the three critical battleground states of the 2004 election could have been due to chance or random error,” he said.

Before 2004, exit polling was considered so reliable that the United Nations used it to police elections in Third World banana republics. Today they’re rarely discussed here — beyond the conventional wisdom that “they’re just so unreliable.”

We didn’t protect the brand

Fifteen years later, I have two additional confessions:

  1. I’m satisfied that Russian interference in the 2016 election, regardless of whether it included outright vote fraud, was the decisive factor in Trump’s victory.
  2. Given the polling data since 2017, there is nothing — absolutely NOTHING — you could tell me that would EVER convince me that a 2020 Trump win would be a legitimate result.

Look, folks, I’m not happy to confess this, and I don’t say that merely because it makes me look like a nut (which it does), and I don’t want to be considered nutty. I say it because my lack of faith in our democratic institutions suggests they’re under increasing stress from the GOP.

No republic — no democratic society, for that matter — can long survive the degradation of it institutions. Faith in each other, faith in the law, rests upon such foundations, and once we begin to doubt them, we’re in trouble.

American democracy is a shambles at the moment because we’ve too long accepted pablum as truth, narrative as fact, absurdity as cleverness. In doing so, we’ve failed to “protect the brand” of American democracy. It’s not worthless, yet, but metaphorically speaking, it’s currently available on a clearance rack at Dollar General.

Run the numbers

Donald Trump is, by every available measure, the most unpopular American president since the dawn of public opinion polling. No candidate this unpopular — incumbent or challenger — has ever won a Presidential election.

True, Trump has some electoral college advantages, but look at the battleground data: He’s underwater in most of the swing states, and not just in terms of approval rating: In recent head-to-head polls, he trails all the leading Democratic contenders. By like 7 to 13 points.

I’m going to pause and stipulate all the necessary caveats about the American electorate and the limits of polling right here — I’m looking at you, Dad — so we can avoid reciting all those objections again. The point isn’t to argue that Trump couldn’t win — he certainly can. And if the voting is within the margin of error, a Trump “win” becomes more likely — because he and his GOP cronies will surely endeavor to steal it.

I expect Trump to lose, but I’m not going to pretend that I would accept a Trump win as legitimate. Like the 2004 exit polls, it simply doesn’t add up. I fully expect him to cheat — win or lose — and my trust in our elections “system,” our media and American society in general is now so threadbare that if he’s declared the winner next November, I simply won’t believe it. Full stop.

If we win, I hope we’ll take this seriously by moving rapidly to ensure the right to vote, accountability in elections, and whatever else needs done to protect our system from being hacked by enemies foreign and domestic. Because for all the things we need to fix now — basically everything, if you think about it — nothing will work right if we’re less than confident in the legitimacy of the process.

Can we all agree on that?

Of course not. We’re Americans.