^{From Page 14 of the WH 2026 Budget Summary}

“States should be responsible for funding their own water infrastructure.” sounds like a very principled statement, right out of old-timey Classic Liberalism, as if it might have been said by the Founding Fathers in a 1787 debate. But the exact words also have a slightly chiding tone, like “It’s time for you states to grow up and take care of yourselves, the federal government can’t do everything.” Either way, from principle or practical realism, the message is conveyed as a stern truth that the Trump Administration wants everyone to acknowledge.

But then there’s this:

Okay, this comes from a ChatGPT ‘discussion’ of a few weeks ago. But the $3.9 billion in annual subsidy for tax-exempt bonds issued to finance water projects sounds about right in my experience. More importantly, the data used for the calculation looks solidly sourced from two major water sector advocacy groups, AMWA and NACWA, who are consistent proponents of tax-exempt muni bonds.

As we know, the muni market was mighty concerned a few weeks ago that their slice of the federal pie (about $25 billion each year in total) would get reduced in the BB Bill as it went through the House. The tax-exemption did get partially whacked in Trump’s 2017 bill, so the fear was not unfounded, and the market’s lobbyists went Code Red, all-hands-on-deck, apocalypse now, etc. Their pushback succeeded in the House; I haven’t heard of any alarums raised yet in the Senate process. More importantly, the WH seems to accept the House bill, which implicitly endorses leaving the tax-exemption untouched. I’m guessing that the muni market will get through unscathed — but there’s still a long and rocky road ahead, so who knows?

What can be done at this point, however, is to ask some hard questions about where the Trump Administration is coming from on federal subsidies for water infrastructure finance. The two subsidies discussed above are obviously different in form, but substantively quite similar — about $2.5 billion per year has gone to the states for new SRF capitalization that can be used for low-interest loans many times over, and about $3.9 billion per year has gone to high-income taxpayers to subsidize about ten times that amount in lower-interest bonds. The same federal principle and/or realism ought logically to apply to both. But the FY 2026 Budget and the WH’s apparent acceptance of the House BB Bill aren’t consistent when it comes to funding (or not) for federal water infrastructure finance.

• Well, Trump Administration, what’s it going to be? Did the statement in the 2026 Budget Summary about states being responsible for their own water infrastructure funding mean anything? Or was it just rhetoric?

• If it’s all just rhetoric, okay, people get that. But what are the realpolitik factors at play here that would explain the Administration’s proposed radical SRF funding cut but apparent silence on the tax-exemption? Lobbyists for both sides can make the same arguments about the critical need for infrastructure funding — you know, the vivid vignettes of children going without drinking water and all that kind of thing. Is it just that the muni bond lobbyists are more badass — that they have better threats? Trump 1.0 took on this crew in 2017 and prevailed — what’s changed?

• Or is something a little more ideological going on in Trump 2.0? As in the standard neoliberal distinction — strict small government principles and/or realism applied to public-sector spending, but a generous hand with ordained private-sector areas, especially major debt markets with high-income investors? That would go some way in explaining another logical inconsistency in the WH 2026 Budget: In the Summary, the SRFs are said to be ‘duplicative’ of WIFIA and, as such, undeserving of any more federal funding. Yet in the full, Technical Supplementary Appendix, WIFIA itself is cut for FY 2026 to just its operating budget, the explanation being that the program can use leftover funding for new loans. Okay, maybe — but isn’t there going to be a big increase in demand from the now lower-funded SRFs? That higher demand will be continuing — will WIFIA be funded again in FY 2027? Or FY 2028? Or…let’s be honest — if the Administration is adhering to the above neoliberal ‘principles’, they simply don’t care, right?

I don’t know when this FY 2026 EPA Budget in Brief was published — presumably right after the Technical Summary was released last Friday? Over the weekend? In any case, despite pretty much continuous searching, I just found it this morning. On page 48, the Brief clarifies something I thought might be the case (but erroneously discounted) in a prior post, specifically in Note 1.

It looks like the Trump Admin does not intend to shut down the WIFIA program, or is not giving an indication to that effect, anyway. Apparently, it was judged that the program could continue loan origination in FY 2026 with ‘substantial’ leftover funding. The rest of the descriptive language is not negative and would seem to acknowledge some success (‘large loan pipeline’), so it seems the cut is simply a way to shave $65m off this year’s budget.

Okay — that’s much better. However, there’s a bit of cautionary signal here — the federal budget is being examined critically, down to the smallest details. For FY 2026 budget, WIFIA’s leftover funding meant that the current goal (savings) could be achieved, apparently without much further thought, and the couch change could be thrown in the pot. But once the microscope is brought to bear, its use might become more standard — and harder questions asked. Like in FY 2027 or FY 2028, when the leftovers are gone and the federal budget is still in parlous shape? It might be better to see this, not as an affirmation of the role or necessity of federal infrastructure finance, but as a stay of execution during which a strong defense can be made — for the next time.

As expected from the earlier budget summary, CWIFP’s proposed FY 2026 budget is just flatlined — no new resources.

I’d like to see how the transportation loan program, TIFIA, is faring. It wasn’t mentioned in the budget summary. But to be honest, I can’t make head or tail of what’s in the Technical Supplement. Apparently, I’m not the only one who can’t tell what TIFIA gets — this from an Eno Center March 2025 article, The FY25 Year-Long CR: What’s in It, and What’s Not:

The $50 million for the formula bridge program at FHWA will once again be matched by $200 million transferred from the TIFIA contract authority program to the same bridge program, as happened in 2024. But between then and now, the 2024 WRDA bill transferred an additional $1.8 billion in TIFIA contract authority to the regular federal-aid highway formulas. That adds up to $2.2 billion transferred out of the TIFIA program in just over a year. No one outside DOT is quite sure exactly how much TIFIA funding is left. This new $200 million transfer will almost completely eliminate the new funding provided for TIFIA by the IIJA for 2025 ($250 million gross, lopped off by the annual obligation limitation to around $200 million. The list of projects expecting TIFIA loans in 2025 is now long and getting longer.

Perhaps to be clear about resources for federal infrastructure loan programs, you need to be clear about what the programs are meant to accomplish — you know, to put the request in context. As yet, that hasn’t been done, but in all the upheaval, I think it needs to get done soon. Such programs might come in mighty handy in certain economic conditions and even to achieve certain ideological objectives, as in ‘facilitate local funding for local infrastructure’ — that’s sufficiently Project 2025-ish, no? It’d be a shame if they all just got chain-sawed away or hijacked for high-risk P3 project finance. Just saying.

The Technical Supplement to the 2026 Budget was just released. Apart from being over 1,200 pages long, it is indeed technical — not easy to read or understand, even for someone who has interest in these matters. In particular, the mechanics of federal cash budgeting are completely sui generis — plain cash budgeting might have a certain logic, but here everything is mixed with terms and categories that reflect idiosyncratic federal practices. Federal insiders will understand the language; to be fair, they are the primary audience. Outsiders like me can only make educated guesses.

The WIFIA program shows up on pages 953-955. The screenshot above is what I think are the critical numbers. The yellow highlight is alarming. Although the WH Budget Summary of a few weeks ago did not show any cuts to WIFIA, the Technical Supplement appears to do so. WIFIA’s prior appropriation level was about $72m — $65m for loan subsidy and $8m for operations. The $8m number proposed for FYE 2026 indicates that the program can keep operating, but there won’t be any new appropriations for loan subsidy.

The brief accompanying text is consistent with this interpretation — it pretty much says in lawyer-like language that while the continued operation of WIFIA is a statutory requirement, that’s all the program will be resourced for. In effect, no new loans: [1]

This appropriation supports all activities necessary for the implementation of the Water Infrastructure Finance and Innovation program established by the Water Resources Reform and Development Act of 2014….The $8 million request to implement the Water Infrastructure Finance and Innovation Act (WIFIA) program is for the Environmental Protection Agency’s (EPA) management and operation of the program, including contract support and associated payroll. The WIFIA program will be administered by EPA’s Office of Water.

Since WIFIA’s borrowers have very high credit quality, $65m of loan subsidy would support about $6 billion of financing — so the impact for the water sector is material.

The importance of the purple highlight is less obvious, but may actually be more concerning. Of course, the startling $1bn of mandatory appropriations for is for interest rate re-estimates, as long predicted on this site — here’s my last estimate, The Economic Cost of WIFIA’s Portfolio at FYE 2023. You can see that the sum of mandatory appropriations for FYE 2024 and FYE 2025 in the Supplement is about $1.6bn, which is in the ballpark of my estimates, though fortunately at the lower end. [2]

The concern is this: The fact that Trump Administration people focused enough on WIFIA to propose cutting $70-odd million from a successful program that finances water infrastructure means that they might also notice the $1.6-odd billion in mandatory appropriations the program has required. Ok, those don’t show up in the budget, so maybe most of the practical folks won’t care — but what about the more ideological ones? Or a technically curious DOGE boy looking to make a mark? And what about muni bond lobbyists who have more reason than ever to put potential competition in a bad light and show how necessary (and relatively efficient) their market is for delivering federally subsidized finance? [3]

I know the budget numbers will change, so it’s not likely that WIFIA cuts, if there are any, will be so drastic. It’s hard to predict where the re-estimate issue will go — there may be ways to manage the ‘narrative’ if necessary. Still — a defensive mindset might be called for.

_____________________________________________________________________________________________

Notes

[1] It is possible that WIFIA has sufficient unallocated leftover subsidy to meet loan demand for FYE 2026 without needing further funding. But I somewhat doubt it — WIFIA usually gets more than enough applications to utilize its annual appropriated funding. There are about 69 ‘pending’ loans on their website — amounts aren’t given, but I’d guess the total is somewhere around $8-10 billion, which would require $80-100 million in subsidy, i.e., all of last year’s funding plus some more. Maybe the ‘unobligated balances’ line above holds a clue? Above my pay grade.

[2] I doubt anyone will dig this deep, but if you include the mandatory appropriations for FYE 2021-2023, the total to date is actually closer to $2.4 billion. The portfolio is now about 50% drawn (i.e. prior loan commitments that were drawn down into funded loans, which triggers the re-estimate), so the final total could go yet higher when the remaining old commitments are drawn, but perhaps by not so much. The actual number is not so important — it’s the ‘billion’ handle showing up in the current budget environment that might be a concern.

[3] There’s this Congressional letter from 2020 re municipal bond market support (what eventually became the Municipal Liquidity Facility). FWIW, note the signatory on page 3, as one of forty-three House members.

^{An AI-generated image of a Neo-Luddite rejecting the AI-based Neo-Narrative}

In prior posts, I’ve pointed out that those who develop and manage narratives, the Narrativators, are highly motivated, presumably very skilled, and likely hard at work on the neo-narrative in preparation for (among other things) the 2026 midterms. This crew will doubtless have the very latest AI multi-media tools at their disposal, which may well enable them to develop something far more powerful than any propaganda ever launched on a population. The stakes for the continuation of neoliberalism are existentially high, the funding sources infinitely deep, and the lessons of 2024 painfully vivid — how could they not succeed this time?

Well, they might not. The chance of that outcome may not be high — who knows? — but it strikes me that the possible origins of such a failure are worth brief consideration, if only as a marker.

The biggest and simplest factor could be that my assumptions about the Narrativators are wrong, specifically regarding their skills and ability to learn from mistakes. Narrative propaganda has, I think, been a rapidly innovating field since around 2008 and even ‘experienced’ Narrativators are likely experimenting a lot of the time. ‘Old’ skills might not be especially relevant; amazing new tools like AI will unleash feverish dreams of innovative approaches that are ‘sure to work’. Likewise, the fact of past failure might be obvious, but the lessons for going forward aren’t so clear. Especially because political mistakes aren’t exactly the same as neoliberal status quo-threatening mistakes. The Democrats’ mistakes last year were political by definition, and they resulted in the loss of the party’s institutional power. Those specific mistakes, currently being much analyzed and discussed, are presumably what will be avoided in future. In contrast, neoliberalism lost its previously effective camouflage in 2024. Rebuilding political power and reestablishing neoliberal camouflage, which so neatly (and fortuitously?) overlapped in 2008, might now involve very different and possibly inconsistent narratives — what do the Narrativators do then?

In short, the neo-narrative could fail in an old-fashioned way: the competence brought to bear was less, and/or the challenge greater, than was thought. More interesting and right up to date, however, are the ways that the involvement of AI itself in the development of the neo-narrative might contribute, not to success, but to the failure of the approach. Off the bat, I can think of three ways — doubtless there are others:

Sudden and widespread revulsion to AI super-slop: From its ancient beginnings, propaganda ‘narrative’ has always in a way been an intentional form of ‘slop’ — curated hallucinations spinning a plausible, but not actually true story. With AI tools, the hallucinations can be more realistic, more personalized, more pervasive — just more everything. The Narrativators will be sorely tempted to push the dial to ’11’, whereupon the resulting ‘super-slop’ may cause overload, annoyance and ultimately revulsion. People will by then have enough multi-media AI in their lives anyway and won’t be easily amused or distracted by yet more of it coming through their screens. For personal economic and social matters, they’ll turn to their own uncamouflaged lived experience, which will be increasingly disconnected from AI virtual worlds. They’ll make political decisions accordingly, uninfluenced and mostly uninfluenceable by the neo-narrative. [1]

Neo-Luddite reaction and resentment: Ah, the Luddites — that ever-handy meme of early 19th century ignorant deplorables smashing the very machines that were to give their descendants (that’s us!) so much prosperity. The meme’s cautionary tale will of course be instantly applied to any modern-day folks who express fear and loathing about AI’s job-killing potential. Presumably, the ones who don’t wish to be considered ignorant or deplorable will then back off — right? Maybe not — this time might be different, for two reasons. First, even in the early stages of the Industrial Revolution it was clear that growing mechanization would require more, not less, labor. Why? Because up until recently machines were powerful but dumb, and averagely smart people were needed to run them. Now machines can be both powerful and at least averagely smart (usually far more); it isn’t clear at all what will happen, and very well-informed pessimism can be justified. [2]

Second, and more importantly for those folks who would otherwise fear the Luddite label, AI is coming for white collar jobs, big time. They may on one level be persuaded that their personal job loss is just part of society’s transition to a glorious future, but on a visceral level, there’ll be old-fashioned, where’s-my-hammer-type anger and resentment. Perhaps the latter will frequently be secretive and partially suppressed, but it’ll likely be a lurking influence on privately made decisions.

Neo-Luddism will undermine any neo-narrative that includes the ‘wonderfully positive transformative’ impact that AI will have on society and our collective material well-being. (Yes, looking at you, ‘Abundance’ proponents.) Less directly, to the extent that the use of AI multi-media is detected in the selling of any neo-narrative (whether pitching AI futurism or not), lurking neo-Luddite anger will help trigger or at least add to the revulsion described in the first point — there’ll be a lot of ‘touchy’ people out there.

AI tools used to expose the truth: It is conceivable that the use of AI tools becomes sufficiently widespread that any propaganda, much less narrative, depending on a poorly informed population becomes ineffective. Yes, I know — it’s unrealistic to think that there’ll be a sudden thirst for truth and knowledge, especially given the state of US education. But AI might enable more individuals to package and distribute more ‘truth adjacent’ or ideological material than before, and that might be influential on the margin. Perhaps the role of independent media (enabled by new social media tools) in the 2024 incineration of the old woke narrative is indicative — more like that, but with AI tools?

This possible path of AI usage development has specific relevance to federal infrastructure finance reform, as discussed in prior posts. Could even be a central factor in deep reform of such a technical area. So, I’ll be coming back to that topic often in future.

__________________________________________________________________________________________

Notes

[1] In historical terms, I think this type of ‘overnight rejection and revulsion’ has precedents for happening rather quickly — from ‘divine right’ to the guillotine in a few years? Or the pre-Reformation ‘indulgence narrative’ (complete with advertising jingles!) triggering a reaction that soon led to widespread (and actively destructive) iconoclasm. A more specific example for neoliberals who feel that they’ll be covered by a narrative, one way or another, learning about the fate of French Pre-Revolutionary tax farmers would be…edifying. Just saying.

[2] One big population did lose out quickly and permanently in the Industrial Revolution — horses. Powerful, flexible and (in terms of their tasks) smart animals — but not smart enough. That population would have had ample reason to be pessimistic in the 19th century. Yes, overall wealth would clearly increase. But other than a small minority (racers, show, pets, etc.), they weren’t going to get any of it.