Saturday, October 24, 2026

The annual deficit will be $5 trillion per year by 2034

 can you tell me the government deficit for the last 5 years, for fiscal year 2022 through fiscal year 2026 (which ended September 30, 2026)?

Yes—here are the U.S. federal budget deficits for fiscal years 2022–2026, in nominal dollars, rounded to the nearest billion:

Fiscal year ended September 30Deficit
2022$1,374 billion ($1.374 trillion)
2023$1,687 billion ($1.687 trillion)
2024$1,815 billion ($1.815 trillion)
2025$1,775 billion ($1.775 trillion)
2026$1,992 billion ($1.992 trillion)

The 2022–2025 figures reflect the updated historical series published by the Federal Reserve using OMB data; earlier reports may show slightly different amounts because of revisions. The 2026 figure is Treasury’s reported year-end result. St. Louis Fed

The deficit increased by approximately $217 billion, or 12.2%, in FY2026 compared with FY2025. A deficit is the amount spending exceeded revenue during that year, rather than the total outstanding national debt.

 Now project the deficits out through fiscal year 2035 assuming that they will increase 12.2% every year. I know there are a lot of other projections and factors that go into this but I am just looking at the recent trend.

Starting with the FY2026 deficit of $1.992 trillion, and increasing it by 12.2% each year, the projection is:

Fiscal yearProjected annual deficit
2026 — starting amount$1.992 trillion
2027$2.235 trillion
2028$2.508 trillion
2029$2.814 trillion
2030$3.157 trillion
2031$3.542 trillion
2032$3.974 trillion
2033$4.459 trillion
2034$5.003 trillion
2035$5.613 trillion

Formula: Annual deficit = $1.992 trillion × 1.122^(year − 2026).

Under this assumption, the annual deficit would roughly double in six years and reach 2.82 times its FY2026 level by FY2035. This is a compounded trend projection using your assumption.

Friday, October 16, 2026

Will Trump blow everything up if Republicans lose in November?

 First read this: https://paulkrugman.substack.com/p/nerobefehl-nation

At home, Trump officials suddenly cut down hundreds of trees at a public golf course in Washington. Trump has long wanted to remake that golf course to his own taste, but that’s unlikely to happen given the political winds, and in any case there was no need to rush the destruction. This was almost surely an act of spite, of deliberate act vandalism, undertaken precisely because Trump knows he won’t get his way. And while a few hundred trees are not in themselves that big a deal, they represent the shape of things to come. Think of the Kennedy Center, a national icon that Trump is destroying because he can’t put his own name on it, as the model for what is all too likely to happen to America over the next two years.

I dismissed this as the ramblings of a mad man on a partisan rant, considering it is coming from Paul Krugman.

But read this:

Recently, President Trump was asked about the upcoming midterm elections. And his answer was shocking: "I think if the Democrats get in, we will end up in a financial depression. We will go into a depression the likes of which we've never seen before. It'll be 1929."  Source: https://www.zerohedge.com/sponsored-post/trump-itll-be-1929

So maybe Krugman is right and Trump confirmed it - Financial Armaggedon will be here on November 4!  And them Krugman can gloat about how he is right and that the answer is to throw not trillions, but tens of trillions of dollars at the next minor recession.

==========================

So I am speculating on a possible outcome.  There could be a 1929-style depression as Trump says, even if he does not torch the place like Krugman is predicting.  Look at diesel shortages (which will increase the price of everything), combined with a housing crash because of the higher rates, combined with a Las Vegas tourism collapse, combined with a water shortage and throw in an AI crash because of overbuilding. There could be a depression, which would then lead in to a 2021-style recovery in 2029-2030 caused by Keynesian/Krugmanite spending.

Monday, October 12, 2026

Judge Ana Reyes vs 500 trees on a golf course - who will win

 Let me just preface this by saying that if you lose your temper and start screaming at someone that you prove that you have no authority.  The screamer is just seen as a very annoying person, and maybe the screamee will try to momentarily appease the screamer until he can escape from the situation, and then he will resent the screamer.  If she will be obeyed, it will be as a dictator, not as a leader who deserves respect.


See also: https://www.newsweek.com/judge-trump-officials-trees-dc-east-potomac-golf-course-12556145

Saturday, October 10, 2026

Michael Hudson is anti-AI for no reason

 Or something like that.

Can you summarize this interview with Michael Hudson: https://www.unz.com/mhudson/the-ai-crash-is-coming-and-you-will-pay-for-it/

Hudson’s central argument is that AI may become a financial bubble whose losses are ultimately shifted onto the public. In this interview with Radhika Desai, his concern is how AI investment will generate enough income to justify its costs. The Unz Review

Their main points are:

  • Uncertain returns: Huge spending on chips, data centers, and research may exceed what customers will pay.
  • Monopoly ambitions: Hudson argues that American companies seek profits by making businesses and countries dependent on proprietary systems and charging for access.
  • Competition from China: Both speakers portray cheaper, open Chinese models as a threat to that strategy, emphasizing practical applications over promises of superintelligence.
  • Costs to ordinary people: They anticipate higher electricity and water costs, displaced workers, poorer customer service, and expanded surveillance.
  • Public bailouts: They predict governments will invoke national security and competition with China to rescue failing investments. Desai makes the strongest comparison with 2008.
  • Human judgment: Hudson accepts useful automation but doubts that it can substitute for contextual understanding and judgment. The Unz Review

My (ChatGPT) reading: This is a political-economic critique, rather than a demonstrated crash forecast. The interview provides neither a clear timetable nor a detailed financial model establishing that a collapse or bailout is inevitable.

My (the human posting this) thoughts, based on not having read the interview notes.  Hudson has not proven either of his points that 1) an AI crash is coming or that 2) you will pay for it.  He doesn't say why it will crash - too expensive, too resource-hungry, or dependent on a debt bubble that pops.  He doesn't say why you will pay for it, although it is possible that Anthropic or other AI companies will get a government bailout.  But why would the government do that, unless there was a systemic threat to the economic system, like in 2008,  which seems unlikely.

AI is just the next level of computing.  An LLM can run locally on a powerful enough server.  It doesn't necessarily require a data center.  Michael Hudson just loves to hear himself talk, and boy can he talk, but he is just spouting his opinions which he doesn't back up with facts.

Ok maybe I was being too harsh on him for now reason and I should read or watch the interview.

Friday, October 9, 2026

Trendline of Deficits

 


Image source: https://uk.finance.yahoo.com/news/wall-street-sees-ominous-sign-090000374.html

The text says: "Fiscal gap has trended around 6% of GDP in recent years".  

That's not what I see - I see deficits that are increasing as a percent of GDP that are currently about 6% of GDP and continue to increase as a % of GDP in an unbroken trendline since the year 2000.

Can you explain Department of Education accounting?

 I think there is some funky accounting going on at the Department of Education.  This is DoD-type obfustication (spell-checker says that word should be obfuscation but I like my spelling).

The latest CBO Monthly Budget Review has this information about rising costs in the Department of Education: Outlays of the Department of Education rose by $41 billion (or 117 percent), driven largely by differences in the amounts the Administration recorded for the estimated costs of outstanding student loans. In September 2025, a reduction of $131 billion was recorded because of program modifications authorized in the 2025 reconciliation act that offset most of the department’s fiscal year 2025 spending. In 2026, the Administration recorded a net decrease in the estimated costs of outstanding loans, mostly stemming from updated assumptions about borrowers’ repayments. That reduction was significantly smaller than the one recorded in 2025, resulting in a year-over-year net increase in outlays. Partially offsetting that net increase for 2026, spending from the Education Stabilization Fund fell by $12 billion and outlays for other components of the department’s budget fell by a similar amount. My question: I don't understand this at all? I thought the monthly budget review was on the cash basis, so the outlays from Education would be disbursements on the loans to schools less payments from borrowers. And of course they measure the amount of student loans outstanding which is somewhere around $7 trillion, but that is an asset; and if they write off some loans that has no effect on cash. But this makes it sound like it is on the accrual basis, which is different from the way the rest of the federal budget is calculated. Can you explain what is going on here?

Your understanding is essentially correct: the federal budget is mostly on a cash basis, but federal loan programs are a major exception. Student loans use an accrual-like method required by the Federal Credit Reform Act of 1990. Consequently, “outlays” in the Monthly Budget Review do not always mean cash paid outside the government. Congressional Budget Office

That exception explains the passage you quoted.

1. What gets counted as student-loan spending?

Instead of counting loan disbursements minus borrowers’ payments as budget outlays each year, the government records the estimated lifetime net cost of the loans, expressed in present-value dollars. That incorporates expected principal and interest collections, defaults, forgiveness, and other repayment provisions. Administrative expenses, such as servicing costs, are counted separately on a cash basis. Congressional Budget Office

Here is a simplified example, excluding administrative expenses:

EventActual external cash movementEffect on reported budget outlays
Government makes $100 billion of loans and estimates the present value of future collections at $80 billion$100 billion paid out$20 billion subsidy cost
Later, revised assumptions raise expected collections’ present value by $10 billionNo immediate external cash movementNegative $10 billion adjustment
Alternatively, revised assumptions lower expected collections’ present value by $10 billionNo immediate external cash movementPositive $10 billion adjustment

Those later adjustments are called subsidy reestimates. Changes in loan terms can also produce adjustments to previously recorded costs. Congressional Budget Office

2. Where do the actual loan disbursements and repayments go?

The government maintains two distinct accounts:

  • A budgetary program account, which records the estimated subsidy costs and subsequent adjustments.
  • A nonbudgetary financing account, which records actual loan disbursements and borrowers’ repayments.

The financing account’s cash flows affect Treasury’s borrowing needs, but they are excluded from the reported budget deficit. The subsidy entries involve transfers between government accounts rather than payments to schools or borrowers. Congressional Budget Office

So your proposed calculation—cash disbursements less borrower payments—is meaningful as a measure of cash requirements, but it is not the measure used for student-loan outlays in the budget.

3. What does the CBO passage mean?

Reading the paragraph you supplied:

  • 2025: Changes authorized by the reconciliation act reduced the estimated lifetime cost of existing loans by $131 billion. That was recorded as negative spending, substantially lowering Education’s reported outlays.
  • 2026: Revised repayment assumptions also reduced estimated loan costs, but by a much smaller amount.
  • Therefore, 2026 reported outlays increased compared with 2025 because the accounting credit was smaller—even though both years included reductions in estimated loan costs.

For illustration only, if other spending stayed at $150 billion, a negative $131 billion adjustment would leave $19 billion in reported outlays. A negative $60 billion adjustment the next year would leave $90 billion. Reported outlays would rise $71 billion without any increase in that other spending.

The actual passage also says ordinary spending declined in two categories, partially offsetting the increase caused by the loan accounting adjustments. The $41 billion increase therefore does not establish that Education paid out $41 billion more cash.

4. Your point about the asset and write-offs is right

Writing down a loan asset does not itself use cash. Under credit reform, however, a reduction in expected collections can increase the reported deficit immediately; the additional borrowing need emerges later as those expected payments fail to arrive. An actual write-off already anticipated in the estimates need not generate another expense. Congressional Budget Office

One numerical correction: outstanding federal student loans are about $1.7 trillion, rather than $7 trillion; the September 2026 Federal Student Aid update puts the federally managed portion above $1.64 trillion. Knowledge Center

You spotted a real accounting distinction: the reported federal deficit combines mostly cash-based spending with present-value accounting for credit programs. It is not a pure cash-flow statement.

What is the current balance of the outstanding federal student loans, and the amount they expect to collect if it is different, and how did that change from the previous year? Sorry for giving you the wrong number.

The latest published balance is more than $1.7 trillion. The latest financial-statement estimate of the government’s loan asset is about $1.29 trillion—but those figures have different dates and coverage.

There are two useful comparisons.

Latest outstanding balances, as of June 30, 2026

Loan portfolioJune 2025June 2026Year-over-year change
All federal student loans, including some held by private lenders$1.67 trillionMore than $1.7 trillionNearly 4% increase, according to FSA
Federally managed portfolioAbout $1.58 trillionMore than $1.64 trillionRoughly $60 billion increase, using rounded figures

These are balances owed by borrowers, including accrued interest. The June 2026 figures were published September 22, 2026. Knowledge Center

Latest comparable financial-statement values

For the amount the government expects to recover, the closest published measure is “loans receivable, net.” The latest annual financial statements I found are for September 30, 2025:

Education Department loan receivables, billions of dollarsSept. 30, 2024Sept. 30, 2025Change
Principal owed$1,445.5$1,473.7+$28.2
Accrued interest owed128.5161.9+33.4
Gross amount owed1,574.01,635.6+$61.6
Less: allowance for subsidy(493.0)(343.3)Allowance decreased $149.6
Net loan asset—present value of expected future net cash flows$1,081.1$1,292.3+$211.2, or 19.5%

These figures cover loans recorded as Education Department receivables, rather than the entire federal student-loan portfolio. ed.gov

The $1.29 trillion is not a prediction that borrowers will pay exactly $1.29 trillion in total. It discounts projected future cash flows into today’s dollars, including future interest collections and the effects of defaults, forgiveness, and repayment terms. Therefore, the $343 billion difference is not simply an estimate of unpaid principal or loans that will be forgiven. The financial statements do not give an equivalent undiscounted total of expected collections in that table. ed.gov

The striking change is that the amount owed rose about $62 billion, while the estimated net asset rose $211 billion. Much of that larger increase came from accounting changes: the 2025 reconciliation reports approximately $147 billion of loan modifications increasing net receivables, plus about $29 billion from net downward cost reestimates. New lending, payments, and other adjustments account for the rest. Those are broader annual financial-statement amounts, so they should not be equated directly with the specific $131 billion adjustment cited in your CBO passage. ed.gov

Thus, the latest data show both a growing outstanding balance and, in the latest annual valuation, a substantially higher estimated recovery value. I would not compare the June 2026 balance directly with the September 2025 net asset to calculate a current expected-loss percentage.

How did ending the SAVE plan effect this? Maybe it didn't change anything currently because it is still being implemented, but for next year, how much is the expected present value of the loans expected to be vs the expected present value if SAVE was not implemented? In other words, how much more money is ending the SAVE plan expected to bring in in FY 2027?

Ending SAVE increased the estimated value of future loan repayments, and that effect was already substantially recorded in FY 2025—even though borrowers’ transition to other plans came later. But I did not find an official estimate isolating additional cash collected in FY 2027 specifically because SAVE ended.

I’m interpreting your comparison as SAVE continuing versus SAVE being ended and replaced by the available repayment plans.

The closest official estimate I found is in the Education Department’s 2026 repayment regulations:

Scope of estimateEstimated reduction in lifetime costs, in present-value dollars
Repayment-plan changes affecting existing loans originated in 1994–2025$121.8 billion
Repayment-plan changes affecting future loans originated in 2026–2035$246.5 billion
Combined$368.3 billion

The comparison baseline included SAVE. However, these figures cover the broader repayment overhaul, including RAP, the new standard plan, and changes to other repayment plans. They do not isolate SAVE’s elimination. The future-loan estimate also incorporates interactions with loan limits. public-inspection.federalregister.gov

For your question about the existing loan asset, the most relevant number is therefore approximately $122 billion more in present-value recoveries from the repayment changes as a package—not $122 billion more collected during one year.

That helps explain the increase in the loan asset I described previously. The Department’s FY 2025 financial report identifies a $131.3 billion downward cost modification from the reconciliation law, covering repayment changes and changes to borrower-defense and closed-school discharge regulations. With an associated adjustment, it reports $147.9 billion in savings from that modification. So neither the $131 billion nor the $148 billion figure should be described as SAVE-only savings. ed.gov

Implementation does not have to finish before the present-value effect is recorded. Credit accounting recognizes changes to outstanding loans when the modifying legislation is enacted, based on projected future repayments. Consequently, FY 2027 collections can increase as borrowers resume payments, while the main lifetime accounting benefit has already appeared in FY 2025. Congressional Budget Office

There are also two different cash comparisons:

  • Against SAVE operating normally: Extra collections would reflect higher payments under replacement plans and less eventual forgiveness.
  • Against SAVE borrowers remaining in payment forbearance: Resuming payments produces additional cash even before considering differences between repayment formulas.

Those comparisons would produce different FY 2027 numbers.

The answer I can substantiate is about $122 billion in increased lifetime present value for existing loans from the broader repayment overhaul. I cannot substantiate a SAVE-only FY 2027 cash figure, or a pair of FY 2027 portfolio valuations “with SAVE” and “without SAVE,” from the official material I found. It would also be incorrect to add that $122 billion to the previously reported $1.29 trillion asset: that asset already reflects the enacted changes.

Thursday, October 8, 2026

Deficit of $28 billion in September

 The federal government incurred a deficit of $28 billion in September 2026, CBO estimates, compared with a surplus of $198 billion recorded in September 2025—a difference of $226 billion. Revenues were $14 billion more this September than they were last September; outlays increased by $240 billion.

The federal budget deficit totaled $2.0 trillion [actually $1,993 million] in fiscal year 2026, the Congressional Budget Office estimates, $218 billion more than the deficit recorded in fiscal year 2025. 

https://www.cbo.gov/system/files/2026-10/61985-MBR.pdf

The debt owed to the public as of 9/30/26 was $32.425 trillion, and the GDP as of 6/30/26 was $32.5 trillion, so the debt to GDP ratio is 99.8%.  The GDP numbers for the 3rd quarter of 2026 will be released in about a month and I will update this ratio then.  

This ratio was 100% a year ago, based on GDP of $30.3 trillion and debt to the public of $30.3 trillion so it really hasn't changed.

The debt to the public increased by about $2.15 trillion between 9/30/25 and 9/30/26, which is about a 7.1% increase.

My projections are for the debt to the public to increase about $2.1 to $2.2 trillion for the next 4 years, with bigger deficits starting about 2031.  The big danger on the horizon is for the Social Security Trust Fund to be depleted about 2033, but I think Congress will just pay the shortfall out of the general fund instead of actually cutting benefits.

=========

Why were outlays so much higher this year than last? The 3 biggest percent increases were: Department of Education (up 117%); Interest (up 11%); Dept of Veterans Affairs (up 10%).

Why did the Dept of Education spending increase so much?  That should be the topic of another post.

The Net Interest on the Public Debt was 1,143 billion in FY 2026, up 115 billion from FY 2026, an increase of 11%.  

If this continues to increase at 11% per year, when will it exceed $2 trillion per year?  That would be in 2032, about the time that Social Security reform will be very important.  So the deficit will really skyrocket after 2032, from the relatively tame $2 trillion per year now, to over $3 trillion per year in 2033 and increasing at least $200 billion per year.