Since my initial rundown of the debt ceiling deal was published, we got the actual text of the agreement. Let’s go through a few particulars that are new:

TANF changes are really a five-state pilot. Most states are not going to be affected. There is a "small checks" scheme where states get families eligible for TANF using their own expenditures. That faces a crackdown, but not until 2025.

SNAP work requirements have a phase-in. It’s a very odd phase-in. Able-bodied, childless adults must work to avoid time limits on nutrition assistance. The cutoff moves from 49 years old to 50 within 90 days of passage (approximately September 1), and then to 52 at the beginning of fiscal year 2024 (October 1). One month for a phase-in? That won’t cause trouble! The full change to 54 kicks in on October 1 of 2024, and then sunsets October 1, 2030.

Permitting really is a nothingburger. The BUILDER Act, which mostly writes down things the White House has already implemented, like a single agency lead, better coordination, and a handful of page limits and timelines for environmental reviews (though they can be extended), is the main thing in here. A provision on interregional transmission lines is not only a study, it’s an extended study that can take a year and a half to write and another year to finalize after public comment. It’s very clear that the monopoly utilities lobbied that provision into dust. There is an addition of energy storage projects to the FAST Act, an old infrastructure streamlining bill from 2008.

They added Joe Manchin’s Mountain Valley Pipeline approval. The Pipeline Payoff (a new term I’ve coined, how do you like it?) is in the deal, confirming outstanding approvals and stripping jurisdiction from all courts except the D.C. Circuit, which is thought to be friendly. Some have taken this to mean that progressives were wrong to oppose last year’s permitting bill, since Manchin got what he wanted anyway without any trade-offs. This neglects that it was Mitch McConnell who torpedoed permitting because he thought he could get something better in the next Congress. He was right. Plus, the Pipeline Payoff is going to drain Manchin’s support for a permitting deal that gives something to both sides. See Dave Roberts on this.

AI environmental reviews? There’s a study commissioned on the potential of "online and digital technologies to address delays in reviews." This is mainly about online portals as repositories of information and better tools to track progress, but it explicitly says the study is not limited to that. I immediately thought about AI. One way to accelerate the review process is to have ChatGPT write it, I guess. That will probably go as well as AI legal briefs. This bears watching.

I’m no longer worried about administrative PAYGO. This was the provision that codified an executive order that forces federal agencies to offset any mandatory spending created by their regulations by cutting other regulations. If it had teeth, it would be very bad, essentially baking in deregulation. But the agency can exempt rules from this requirement, the Office of Management and Budget can waive it entirely, there’s no judicial review for any OMB finding, and it sunsets at the end of 2024. This may slow down a couple of rules, but I don’t expect it to do much.

The unobligated COVID aid rescissions come to $29 billion. That’s the approximation given by negotiators. In addition …

There’s potentially $21.4 billion in cuts to IRS funds coming. Officially in the bill, the cut is $1.4 billion. But according to numerous reports, budget appropriators plan to steal $10 billion from the $80 billion in IRS funding in the Inflation Reduction Act in fiscal year 2024, plus another $10 billion in FY2025, and use that to fill in gaps in nondefense discretionary spending. (It’s possible those losses get replenished later.) The COVID rescissions also backfill that nondefense discretionary budget in FY2024. That brings us to …

The automatic continuing resolution creates an interesting dilemma. As I explained over the weekend, the deal only creates topline numbers, baselines for future budget appropriations that have yet to be written. If those bills aren’t passed by January 1, 2024, an automatic continuing resolution snaps in at the level of fiscal year 2023, reduced by 1 percent. (This "penny plan" is the brainchild of libertarian Rep. Thomas Massie (R-KY), who sits on the House Rules Committee and presumably will advance the bill through the committee for that reason.)

But that auto-CR treats defense, veterans, and nondefense spending far differently. Specifically, it’s much harsher on the defense and veterans side. Those buckets of spending were increased significantly in the debt ceiling deal. So if they go all the way back to FY2023 levels, it’s a much deeper cut from where they would have been. Adjusted for inflation, the cut to defense is nearly 10 percent.

Let me show my work. Skip this if you don’t like math; my sources include the CBO baselines and the bill text. Per Bloomberg’s Erik Wasson, FY2023 discretionary spending was $1.602 trillion; the new baseline in the deal for FY2024 is $1.590 trillion. If the spending bills don’t get done, the auto-CR trigger for FY2024 is $1.602 trillion minus 1 percent, which comes to $1.585 trillion. So, a $12 billion cut if the spending bills pass, a $17 billion cut if they don’t.

But under the spending bills passing scenario, defense spending increases by $28 billion, while nondefense is cut by $40 billion. The $29 billion in COVID rescissions and $10 billion from IRS funding (in the first year) mostly fills that nondefense gap.

If the spending bills don’t pass, you get a nominal cut to defense of $9.12 billion, and a cut to nondefense of $7.88 billion. However, the IRS fund transfer, which is not in the deal and is just presumed as part of the appropriation, would not happen under this scenario. (I don’t know what happens to the COVID aid in this scenario; I guess it would just go to deficit reduction.)

OK, you can come back now. The upshot is this: If Congress doesn’t pass the spending bills, the military loses about $37 billion from what they’d otherwise get, a 4 percent nominal cut and adjusted for inflation almost a 10 percent real cut. And, the IRS would keep almost all of its funding boost. The trade-off is that nondefense discretionary would lose around $7 billion more, a little less than 1 percent, in this scenario. Plus, veterans funding would lose a significant amount, probably around 5 to 7 percent.

For additional context, nondefense spending got a major boost in FY2023, which became the new baseline. An inflation-adjusted cut from that is certainly not good, but we’re getting one either way; the difference is less than 1 percent. In the no-spending-bills-pass scenario, you trade a slightly larger cut—off a high baseline—for serious cuts to defense and preservation of the ability to go after rich tax cheats in future years. Isn’t that something requiring serious consideration?

It’s not like a major newsmagazine didn’t just release a blockbuster report on how badly the Pentagon is getting ripped off. Just curtailing routine price-gouging would easily cover this cut, which we should be doing anyway. And even more important, the auto-CR would retain the relationship between defense and nondefense spending that has been severed by this deal. In the future, it’ll be fine to increase military spending without increasing non-military spending concurrently, unless this deal collapses.

The whole thing depends on Congress not being able to pass its 12 appropriations bills. Doesn’t that seem like a good bet? Congress is famously not good at such things. The House canceled a bunch of votes on the easiest spending bills last week because they couldn’t find consensus. The point of this auto-CR is to concentrate minds and get them moving to pass the bills. But if you’re a Democrat who talks about how we have a bloated military budget, and that rich people get away with murder from an imbalanced tax system, isn’t this your big chance? And all you have to do is just not vote for spending bills! Every Democrat who does this makes it that much harder to get those bills passed, and that much closer to kicking in the auto-CR.

If you asked me point-blank, I’d say that Congress will somehow get those bills done. The military has contractors in practically every congressional district, making military pork barrel spending a jobs program for almost all politicians. And the fact that veterans programs would be clobbered too makes this a heavy lift. But to be clear, all that’s needed here to trigger the auto-CR is congressional incompetence. That’s not exactly out of the realm of possibility.