President Trump promised tax reform, so Senate Majority Leader Mitch McConnell and his colleagues are considering a proposal that would hamper the retirement prospects of about 49 million Americans to benefit the heirs of the 5,500 or so estates a year that are currently subject to estate tax. (J. Scott Applewhite/AP)

When I write about the strange ways that Washington math sometimes works, I’m usually sort of playful. That’s because if you’ve dealt with real-world accounting for a long time, as I have, Washington math can be silly enough to make you laugh.

But today I’m afraid that flawed Washington budget math is likely to do serious damage to tens of millions of Americans who are trying to save for their retirements. Given that the number of people earning substantial pensions is shrinking rapidly and Social Security benefits aren’t enough to live on, saving for retirement is essential.

That’s why it’s so appalling to see Congress — okay, many of its Republicans — considering a proposal that would hamper the retirement prospects of about 49 million Americans to benefit the heirs of the 5,500 or so estates a year that are currently subject to estate tax.

This may happen because of Washington’s bizarre budget math.

What has me close to ranting is the serious chance that the tax-cut bill would greatly reduce the tax deductions that about 49 million people now get when they stash money into 401(k) plans and similar tax-deferred retirement accounts.

President Trump tweeted on Monday: “There will be NO change to your 401(k). This has always been a great and popular middle class tax break that works, and it stays!”

Well, maybe. After Republicans, including Chairman Kevin Brady of the tax-writing House Ways and Means Committee, said that shrinking the tax-deductible amount is being actively considered, Trump seems to have backed off at least a little bit, and maybe more.

Here’s the problem I have with the budget math of limiting 401(k) contributions. The money that my employers and I put into my 401(k) plans over decades was tax-deductible at the time and is becoming taxable as I take out my annual “required minimum distributions.”

I suspect that if you reconciled the taxes I’m paying on my distributions with the deductions when the money went in, you’d see that even adjusted for time, the taxes offset most if not all—or maybe more — of the deductions my employers and I got all those years ago.

But those future revenues aren’t counted in budget math when the deductions are made. All that counts is the deductions.

Had I put that money into Roth accounts, where contributions aren’t deductible but whose withdrawals aren’t taxed, the Treasury would have collected more taxes from me then, but would collect nothing now. I’d probably be better off, and the Treasury worse off.

The Republican talk is that people will be able to stash money in Roth-like accounts, so they could still save.

The problem is that a lot of people — including my wife and me back when we lived paycheck to paycheck — would find it difficult, if not impossible, to put money away if it weren’t deductible.

So even though I suspect that a rigorous analysis of swapping 401(k)s for Roths would show the Treasury losing money over the long haul, doing that would help budget numbers look better (or less bad) today. And today’s numbers, not reality, is what matters.

And while we’re in Bizarro Numbers Land, let’s discuss a prediction by someone I like and respect — but whose math seems squirrelly to me: the widely ballyhooed $4,000-per-household annual benefit that Kevin Hassett, chairman of the president’s Council of Economic Advisers, says will be generated by cutting the corporate tax rate to 20 percent from the current 35 percent.

Hassett and I met and became friends in 1999, when he and Jim Glassman wrote a book called "Dow 36,000" and I attacked it.

It was a bestseller, full of plausible-seeming scenarios and lots of econ-speak. But let’s look at what I consider the key passage, on p. 13: “. . . it is impossible to predict how long it will take the market to recognize that Dow 36,000 is perfectly reasonable. It could take ten years or ten weeks. Our own guess is somewhere between three and five years.”

Oops. The Dow was 10,824 when the book was published in September 1999. Three years later, it was 7,986. And 10 years later, 9,779. On Friday, a near-record 23,434. Not remotely 36K.

In a conversation Friday, Hassett told me to start my Dow math with the 8,906 at which it closed on March 30, 1998, the day he and Glassman published a Wall Street Journal op-ed that led to the book. Okay, Kevin, it’s done.

Hassett also said I was misinterpreting “Dow 36,000.”

“What we said was that if you buy equities [stocks] and hold them for the long run, you can do very well,” he told me.

That’s what they said, they’re right. It’s true. But of course, “stocks for the long run” wasn’t the sales pitch — Dow 36,000 was.

Similarly, Hassett and his CEA associates told me that their $4,000-per-household average annual projected gain in salary and wages would take 10 years to be realized. They also told me that the current average of U.S. households is $83,000. So the rise — which I still don’t buy — would be about 0.5 percent a year.

That time period and that low annual projected rate of gain aren’t exactly stressed in Republicans’ sales pitches — but I certainly don’t consider that Hassett’s fault.

The bottom line: If by Thanksgiving we end up with a tax bill whacking 401(k)s and projecting quick tax-related growth in household incomes, we won’t have to wait until Dec. 31 to pick the turkey deal of the year.