May, however, is typically marks the inflection point between tax refund/bonus season portfolio improvement and the deterioration we see the rest of the year through the Holidays. DLQ rates are usually, though not always, a little worse in May with mild DLQ rate rises and often a modest SDLQ drop. My prediction from last month was

So what did May show for the 7.2 million FHA loans in GNMA II MBS this year?

Read more

First, the good…..

The Total Treasury Model (“TTM”) projected a Financing Need of 547b and Marketable Borrowing of 661b for the Jul-Sep qtr.

Treasury projected a Financing Need of 557b and Marketable Borrowing of 671b for the same qtr. That is fairly remarkable alignment this far out. As part of that Marketable Borrowing estimate, Treasury agreed that the most likely TGA level (rounded to the nearest 50b) at the close of business on Sept 30 is 950b, a rise from end of Junes 900b. Note that Im not just guessing that Treasury is going to raise the projected end of qtr balance (which again is not a target). The TTMs detail output tells me this is likely.

The image above is from the detailed TTM output released with last Sundays QRA projection post (the same output is included with every TTM update post). Note that the TGA End level the TTM projected for 9/30 was ~968b (orange highlight). Treasury always lists in 50b increments so I round that down to 950b and use that as my projected end of qtr level that Treasury will use in their QRA projections.

The daily projected TGA end level is an application of the days TGA cashflow (projected Daily Treasury Statement Table II deposits shown in column D minus withdrawals in Column E) the projected debt issuance/redemption shown in columns F and G, any Treasury buyback operations in column H and projected bill drop (the difference between the face value of treasury bills and the cash Treasury actually receives, e.g. Treasury may sell 100b in 4-week treasury bills but only receive 99b in cash for them, thus a 1b bill drop).

I also calculate the projected TGA risk policy level for that day and the projected cushion over it. I figure out likely bill issuance by iteratively running the model and tweaking bill increases/reductions to get the cushion over the risk policy high points mid months and end of month to be just a small amount (10-40b range ideally) for the extent of the model run.

Sometimes, it becomes clear that Treasury would have to dramatically alter bill issuance to hit a TGA risk policy high point. Since Treasury favors more stable and predictable bill issuance patterns, those setups are more likely to be met with a short term cash management bill as a patch to get over the high point.

I predicted one for the end of this month.

which Treasury confirmed they intend to do.

So in general. the TTM did very well this QRA cycle with one glaring exception

Apr-Jun

The TTM predicted a Financing need of -1b for the qtr.

Treasury, which had previously (winter QRA) been projecting an 8b Financing Need increased it by a whopping 114b to 123b.

The end of June is only 2 months away and the huge flows with the most room for error (the April tax receipts) are already in for both the TTM and Treasury’s model. Its a huge amount of short term disagreement which is even more striking when you consider the alignment over Jul-Sep. So what are the likely causes of this disagreement?

Possibility 1. Treasury is much more front loaded with when the tariff refunds will start cash flowing out of the TGA than the TTM has modeled. I mostly relied on the research of others like JPM on the likely schedule for when these refunds would flow out. Perhaps Treasury thinks they will go out much more quickly than what market consensus seemed to be. Through the first week of May there is no evidence of these refunds going out in the Daily Treasury Statement data, but if this is the explanation, I would not expect to start to see it until very late in May through June.

Possibility 2. Treasury is expecting considerably less non-withheld income in June due to the changes in the OBBBA. On further evaluation of the TTMs expectations I think the TTM was expecting 10b more than it should have for paper receipts in June given January’s performance for the non-withheld/Other category. Furthermore, through the first week of May the TTM was projecting about 10b more (almost all of it on the first of May) than actually came in. So it seems this would explain 20b of the 125b difference between Treasury and the TTM. We will just have to see how June plays out.

Possibility 3. Treasury expects Withholding to underperform trend growth over the next two months due to folks getting large tax refunds this year and subsequently adjusting their withholding amounts. Through the first week of May there is no evidence for that with withholding coming in just a smidge heavy vs. the TTMs expectations, but Ill keep watch on it.

Possibility 4. Treasury expects Iran war expenses start hitting in mass before the end of the quarter. I have no doubt there will eventually be a large bill due to Iran conflict expenses that will be paid via cashflow out of the TGA, and to some extent there already has been elevated spending out of the Dept of Defense (DoD) - misc, but those flows are not massive (5-7b/month that is already baked into the TTMs expectations). I also think large expenses like radar installation replacement etc. will take a fair amount of time to contract out and then eventually cashflow out of the TGA. Perhaps Treasury has a different opinion. Through the first week of May though potential Iran war expenses are actually lighter than the TTMs beefed up expectations for them. Ill keep watching

Possibility 5. Treasury expects some wildcard spending item like Federal Financing Bank to make some 50b withdrawal. This occasionally happens (and I am unaware how to project these) but we are pretty far removed from the SVB era cleanup which causes a few of these a few years ago so Im doubtful this will be the case. Anyways Ill keep watch.

Possibility 6. Treasury is just wrong. The TTM has outperformed Treasury’s models for 4 consecutive quarters and a number of them involved Treasury revising their initial estimate the wrong way.

The next two months will be interesting to watch and Ill keep yall posted on how this horse race develops.

Full TTM Output

For paying subscribers, the full updated TTM output through the end of the fiscal year is attached below the paywall.

Read more

Read more

Read more

Greetings:

Treasury has done something interesting over the past two weeks. They kept the bill issuance cuts coming, now 4 weeks in a row. I certainly did not expect (or project) them to do that but it makes sense and it didnt stun me that they did it. In fact, emailing a client two weeks ago I wrote

Treasury seems to be taking a more aggressive stance this year in blunting how high the TGA will spike as a result of the April tax receipts that will start hitting in earnest 10 days from now, with the 3 fifteen billion dollar cash management buybacks and these extended bill cuts. This will have ramifications but Ill discuss those later in the piece.

Total Treasury Model Recent Performance

TTM performance over the past two weeks has been solid enough. In my last post I projected a TGA end level of 768.934b on April 2nd.

The TGA actually closed at 776.076b on April 2nd.

7b off which is not too bad particularly given the higher volatility in some of the larger TGA categories recently (tariffs, tax refunds etc.). If knowing in detail what the TGA is likely to do day by day through the end of the fiscal year (full model output below the paywall) is helpful to your investment analysis, please consider becoming a paid subscriber.

Recent TGA category updates

Every couple of weeks I update the TTM to adjust various deposit and withdrawal category expectations based on what is actually happening in those categories (to the extent it differs from projected) and what is likely to start happening. This cycles updates include

Deposits

DHS - Customs Duties Taxes and Fees (aka tariffs)

March settled at 25b, about halfway in between my initial estimate of 27b made right after the SCOTUS decision and my adjusted 23b based on early receipts in the month. Accordingly I tweaked future expectations to 25b/month. Obviously, this category is super volatile with this administration so i fully expect to update it again, but for now 25b/month seems reasonable.

HHS - Medicare Premiums

These have been up vs model projections a good chunk every month since the turn of the calendar year. I dont exactly know why, but the data tells me this rise is more likely structural than transient so Ive adjusted my expectations for this category upwards through the end of the Fiscal Year.

Withdrawals

Dept of Defense (DoD) - misc

I upped expectations 2 weeks ago in the last TTM update cycle, but overdid it at least for March. To be sure I do expect outlays in this category to increase (vs. historical pattern) through the end of the fiscal year but it seems the outlays may take longer than I thought to hit and thus dialed back a bit of my expectations.

Bill Issuance

With four straight weeks of bill issuance reductions, Treasury is definitely blunting the TGA rise in mid April (it should break 1 trillion, but only briefly), but they are also creating a hole for themselves to dig out of in late May to get over the highpoint in the risk policy level in advance of the June 1st issuance. They have three options to do that: 1. Issue a very sizable cash management bill in mid May maturing in early June to keep the TGA topped up over that span 2. Hike bill issuance rates in early May (2 weeks of 5b increases or 1 week of 10b across the short tenors should about do it) or 3. Some blend of the two.

Presently I have it modelled at 1 week of 5b increases on 5/5 and 5/7 which would require a ~60-75b CMB in mid May, but I am just guessing at how Treasury will solve the problem. What is certain is they will have to do something in May to get them over that May high point with the full magnitude of that something revealed in late April after the receipts become certain. Ill of course keep the TTM updated along the way

Full Model Output

Included in the full model output below the paywall :

1. Daily projections of the start and close level of the TGA each day thru the end of the fiscal year

2. Projections of every deposit and receipt category every day thru the end of the fiscal year

3. Projections of all treasury issuance both announced and yet to be announced. I am am able to closely project the timing and extent of changes to bill issuance

Thanks for your support,

John

Read more

My apologies for the length of time between the last treasury model update and this one. Ill keep them more frequent going forward. Certainly alot has changed in the almost 4 weeks since my last update and I have made a few tweaks to some TGA categories to reflect those changes but mostly the TGA continues to behave as the Treasury model has projected.

Total Treasury Model Recent Performance

Four weeks ago the TTM projected a TGA close balance on March 19th of 886.489b.

The actual close balance on the 19th was 882.822b

Only 3.6b off from nearly a month out. The TTM is not perfect but isnt bad either.

Recent TGA category updates

The projection skill of the TTM is due in part to my regular evaluation and tweaking of all TGA categories where the actuals are straying too much from the projections or categories where some event makes a change in a category’s future flows likely (e.g. the SCOTUS striking own the tariffs or the conflict with Iran). Over the past 4 weeks I have updated the following categories

Deposits

Federal Retirement Thrift Savings Plan - I have written in the past how this category experiences heavy deposit flows when federal workers get nervous about markets declining and shift their retirement assets from equity funds in the TSP to the G Fund in the TSP. Whether that fully explains the heavy deposit flows in this category recently or if there are other structural explanations I dont know but the reality is this category has been experiencing heavier flows of late consistently, accordingly Ive updated the model to expect such heavier flows going forward.

DHS - Customs Duties Taxes and Fees (aka tariffs)

Last month I wrote how I had dialed back tariff expectations from 33b/month to 27b/month. Seems the real number will be closer to 23b/month and I have updated the model accordingly

Non-Withheld Income Tax (Electronic and Other)

The TTM has been doing a fantastic job projecting the higher individual tax refunds this cycle due to the OBBBA changes last summer, but I underestimated their impact on reducing non withheld taxes paid this spring. Ive updated the model to expect 15-20b less in non-withheld income tax through April. Those adjustments have the model tracking well through March now. April will be a swirl of those changes as well as capital gains tax receipts so it will be impossible to separate out but I think the TTMs projections for the non-withheld categories will do well.

Taxes - IRS Collected Estate Gift misc

I changed the modelling methodology on this category. Previously I was using an average of the last 3 months prior to the turn of the fiscal year to determine monthly projections for the next fiscal year. I have updated it to a flat 1b/month projection which fits the realized data for this year much better.

Dept of Education (ED)

Student loan repayment continues to tick up and I adjust the models expectations by about 8% to reflect those higher payments. The model ups that again to 20% in July when the Repayment Assistance Plan comes online but I may have to revisit that depending on whether it looks likely the SAVE plan borrowers will be forced on it when it comes online in June.

Withdrawals

Dept of Defense (DoD) - misc

The immediate TGA impacts from the conflict in Iran are fairly small honestly, though there has definitely been some uptick in this spending category. There will be more uptick as we go later in the fiscal year and new munitions need to be made etc, with much of that outlay coming through this category. Note the proposed 200b funding bill would have a considerably bigger impact than I have currently built in, but I will wait to update the model until the details of that funding bill are finalized with a passed bill.

USDA - Other Farm Service

The TTM wasnt previously accounting for the Farmers Bridge Assistance program which should add ~11b in additional spend in Mar and April. It does now.

Dept of State (DOS)

Withdrawals were running higher so I updated future expectations to better reflect whats actually been happening in this category this fiscal year, especially with the conflict in the Middle East.

Taxes - Business Tax Refunds (EFT)

Business tax refunds were coming in heavier than the TTM projected so it looks like I underestimated the OBBBA impact here as well. I tweaked the expectations for the category up.

HUD - Federal Housing Admin (FHA)

These withdrawals have been coming in notably lower this fiscal year, Accordingly, I reduced the future expectations by ~30%.

TREAS - Claims Judgments and Relief Acts

Not a big category but it was getting well offsides. FY25 had considerably more withdrawals than 24 and 23, but it seems now that FY25 was the anomaly. Modelling expectations based off a blend of FY23 and 24 works much better for this FY and I have updated the model accordingly.

Unclassified

Actual unclassified withdrawals were consistently coming in less than what the TTM was projecting so I cut the full FY expectations by about 10b which now has the monthly projections performing well.

Bill Issuance

Treasury decreased bill issuance across the shorter tenors (4w, 6w, 8w) over the last two weeks generally as expected/projected by the TTM though they did it a week earlier than originally thought. Those reductions are very likely over now and the bill issuance levels will stay the same until one last round of reductions in the first week of June. Starting in July Treasury will need to significantly increase bill issuance rates to fund the deficit heavy months of July and August. The TTM currently has 3 straight weeks of increases starting the second week of July before some reductions in early September before more increases in October at the turn of the fiscal year. The full details for bill issuance are in the attached model output as always.

Full Model Output

Included in the full model output below the paywall :

1. Daily projections of the start and close level of the TGA each day thru the end of the fiscal year

2. Projections of every deposit and receipt category every day thru the end of the fiscal year

3. Projections of all treasury issuance both announced and yet to be announced. I am am able to closely project the timing and extent of changes to bill issuance

Thanks for your support,

John

Read more

I expected that January would be the high water mark for delinquency and serious delinquency rates since partial claims and loan modifications that were started in October could complete their 3 required trial payments for the partial claim/mod, hit the data as “cured” loans and bring the DLQ and SDLQ rates back down. But………. that assumes the borrower was able to successfully make their 3 required payments. What does the February data have to say about that?

Read more

January represents one last month of this mechanical effect. Loss mitigation workouts like partial claims and loan modifications entered into since the new rules took effect on Oct. 1 literally cannot have completed their trial payment plans and thus reflected in the data as current and lower the numerator in the delinquency rate calculations. So January will likely be the high water mark as some of these plans start to complete next month.

Just how high will the delinquency water reach?

*** No paywall this month but please consider becoming a paid subscriber to support my work ***

Summary….

• DLQ is up 22bps in January, which is actually pretty big considering it dropped 65bps last January and 15bps the January before.

• SDLQ is up huge in again in January, nearly another 50bps

• Detected partial claims are slightly down but mostly the same as the last two months suggesting the level of natural cures is ~8500/month

January Partial Claims

My algorithm detected 8401 standalone partial claims were completed in January covering ~43,200 missed payments on those 8401 loans. Down but only slightly from Decembers 8837.

Since partial claims could not have actually been completed in January due to the requirement of 3 successful trial payments. The data suggests that the current level of natural loan cures in a month (where the borrower scrapes together the funds to pay off the arrears and bring the loan current) is actually in the vicinity of 8,500 a month for the FHA loans in MBS.

Headline DLQ significantly Higher in December. SDLQ goes boom again

The MoM headline delinquency rates for FHAs in the GNMA II portfolio (about 7.1/8.1 million total FHAs) saw a significant increase in January rising from 12.17% in December to 12.39% in January.

Serious delinquency rates surged higher again rising from 4.68% in December to 5.15% in January and are now almost 40% higher than where they were in October

The portfolio degradation was again concentrated in the serious delinquency. There was again improvement in the number of loans 1 month late which is consistent with seasonality for this time of year, loans 2 months late also improved. Loans 3 or more months late however all showed continued degradation, which got worse the more delinquent the loan was. This degradation is very much tied to the inability for loss mitigation workouts to “cure” the loan right now because of the requirement of 3 trial payments.

I wrote this in last months update which was a pretty accurate projection of what was going to happen this month.

“Looking to next month…. seasonally there is a tendency for small declines December → January in the 1 month late and 2 months late portions of the portfolio. Regardless, the DLQ rate will likely rise because the partial claims and loan modifications (the ones that actually prove successful) will need another month for Trial Payment Plans to complete before they can start “curing” serious delinquencies in the data again. Serious DLQ will likely see another 45-50bp rise and should end January at over 5% SDLQ.”

February is harder to project. Seasonally there is a tendency for declines in DLQ , particularly in the 1 and 2 month late cohorts through May, before a spike in June through the end of the year. Mostly the same story with SDLQ. But this year is not a usual year since this February marks the first month where trial payment plans can be completed and loss mitigation workouts can reflect in the data (by bringing the loans current) under the new procedures. Industry sources tell me they expect roughly 50% success/fail rates on the TPPs. The February data will give a good read on what the actual rate will be but whatever it ends up being, improvement in both DLQ and SDLQ is expected. How much is the key question.

January FHA GNMA II MBS Performance Charts - Full Portfolio (excluding mods/reperforms so the scale doesn’t skew)

DLQ up significantly , SDLQ up huge. Almost certainly the high water mark for the season, possibly for the cycle.

Modifications and Reperforms

The Modifications of Doom, the modifications performed since mid 2022, again saw worsening performance with no partial claims or loan modifications available to “cure” loans. 54.1% delinquent and 29.3% serious delinquent speaks for itself. As mentioned last month, this performance significantly understates just how bad the performance really has been since many of the modifications that were originated during this period have already been removed from the MBS pools due for delinquency mgmt purposes or further loan mods. When all is said and done, the success/fail rate on these modifications is going to be scandalous (already is). FHA loans are a big program and the losses to the taxpayer here are likely more than offset by profits on the up front and monthly mortgage insurance payments but thats no excuse for the magnitude of the policy mistake on these modifications.

January FHA Recent Originations Charts

Recent originations (between 3 and 15 months ago) also saw rises in DLQ and SDLQ though the DLQ rises were in recent rates refinances and recently originated mods. Outside of recently originated mods (which includes many of the mods of doom) the recently originated mortgages continue to perform better then the full FHA portfolio

January FHA Claim vs. No Partial Claim charts

Broken record alert…… but once again your friendly monthly reminder. Whether a loan has had a partial claim or not in its history makes it 5-7 times more likely to be DLQ or SDLQ and over the last couple of months this disparity is further widening a bit.

While there are a few subplots to the FHA story, the performance disparity prior partial claim vs. not is the main plot by far.

Pool Removals in January

Payoffs dipped which likely partially reflects some seasonality and also the hump in FHA rates in mid November prior to new lows since then. Refinances that rate locked in late November/early December should hit with next months data so I expect a rebound in payoffs then.

Removals due to Foreclosure steady, though still at low levels. This will be the last place we see the impact of the tightened loss mit procedures since foreclosures can take 18 months to make their way through the process in some states.

Delinquency Repurchase pool removals fell a touch but remain very high. I expect these will remain high and actually grow some if prevailing rates dont rise again.

Removals for imminent loss mitigation found their bottom. Next month they should start to rise as loans that completed their 3 trial payments for a modification start being removed from pools to complete the modification. How many? Thats the interesting question. Stay tuned…

Data Behind the Charts

Gnma Portfolio Performance Jan2026

479KB ∙ XLSX file

Download

Download

Thanks for your support,

John

The numbers with the preceding 3 fiscal years of context:

Electronic (EFT) receipts ~+40b over last January:

Other (paper checks) receipts ~-10b over last January:

A net increase of ~30b over last year which supports my posit that capital gains tax receipts will be heavy again this April (though it also indicates that the Jan/Apr percentage share of the overall fiscal year take will probably more closely match 2023 than 2025). I updated the Total Treasury Model (TTM) to now expect a total of 1.29T in non-withheld income taxes for full fiscal year 26. This is split 840b for electronic receipts and 450b of other/paper receipts and represents a 10b overall increase to what the TTM had in mid January.

Other TTM updates

In addition to the non-withheld income tax updates. I updated a number of other categories where the projected amounts were straying from the actual amounts. Some of that stray is due to anomalous receipts/outlays that I just hardcode in as having happened, but some of that stray is structural vs. what I had previously modeled (e.g. non-withheld taxes) and I update those categories in such a way that future expectations are impacted. The list of modified deposit categories are:

The list of modified withdrawal categories are:

I perform these updates every couple of weeks to give the model the best chance to be accurate with its future projections.

Updated QRA call

Jan-Mar is 27b lower on the financing need than my initial call a few weeks ago reflecting in significant part the higher realized January tax receipts.

Apr-Jun is nearly unchanged at 2b higher

Coupons

I neglected to make a call on coupons two weeks ago, mostly because I dont expect Treasury to change them. I would be shocked if they increased them meaningfully (meaning more than some 1b increase to a TIPS or FRN issue), particularly now that the Fed is buying significant amounts of treasury bills which will reduce the marketable borrowing pressure on Treasury further when those bills start maturing in the march to may timeframe. I think it is also useful to look at the projected net issuance of coupons and bills between now and the end of June. (the full model output for paid subscribers shows the projected net issuance on every day treasuries issue/mature).

For coupons that net issuance is ~ 678b. For bills that net issuance is ~ -57b. Granted the Jul-Sep figures are much more balanced at 330b and 317b but I just cant see Treasury increasing it this QRA or the Spring for that matter, maybe Summer.

I think its more likely Treasury would use the Feds purchase of T-bills and the accompanying reduced marketable issuance needs of Treasury bills as potentially a justification to reduce coupons in the short term so they can continue to supply the amount of bills that the market “desires”. I dont project this, but I would be less surprised by this outcome than I would be if Treasury opted to hike coupons.

Full Model Output

Paid subscribers receive the full updated model output every couple of weeks with daily projections of future Daily Treasury Statement content thru the end of FY26 including :

1. Daily projections of the start and close level of the TGA each day

2. Projections of every deposit and receipt category every day

3. Projections of all treasury issuance both announces and yet to be announced. I am am able to closely project the timing and extent of changes to bill issuance

Thanks for your support,

John

Model output is below the paywall.

Read more

That trend of shrinking financing need in FY26 looks likely to continue as the non-withheld tax receipts received via EFT in mid January were strong though we wont know the full extent of the strength until the paper checks come in over the next week and a half.

I recognize that we are in a mid term election year that the current administration is likely particularly driven to win and that often leads to policy responses that involve spending. So perhaps the train speeds up with policy actions not yet taken, but at least for now, it is slowing a bit in the actual data and putting aside potential midyear shenanigans, I expect that to continue through the end of the fiscal year.

Projecting the Winter QRA

The winter QRA is the most challenging to project and Treasury’s error between QRA projected and actuals over recent years bears that out. The reason its tough to project is not due to regular spending. By this point in the fiscal year there is enough data to inform good projections on the big spending categories (Social Security, Medicare, Medicaid etc.). Surprises occasionally occur in these (and the smaller) spending categories, but they often miss in opposite directions vs. whats modeled thus netting out and keeping the models headline financing need projection close to what gets realized.

The challenge is projecting the April tax receipts and projecting the tax refunds which are mostly paid out in the early February to June time frame.

Individual tax refunds have been very consistent over the past 3 years clocking in at 298.5b in FY 23, 298.9b in FY 24, and 302.7b in FY 25. This fiscal year however, they should rise markedly due to tax policy changes in last summers budget reconciliation act (the One Big Beautiful Bill). The act contained significant tax cuts (no tax on tips, SALT increase, auto interest deduction etc.) but the IRS did not adjust the withholding tables after the the OBBA became law. Thus, folks were more likely to overpay on their withholding last year. The tax foundation estimates these income tax cuts at a total of 129 billion. But… assuming they are correct, will all of that actually translate to higher refunds or lower receipts when folks file by April 15th. People are not always the most efficient with their taxes.

Complicating the picture is the other big impactor on April Tax receipts, capital gains taxes. I did a detailed dive into capital gains tax receipts last year in my 2025 Winter QRA projection piece if you are interested in the mechanics behind how I make my projections.

For this year, the setup is largely the same. The April/May tariff drop aside, the S&P 500 rose steadily last year ending the year near all time higs.

same generally with the nasdaq and russell 2000

If you bought and sold a stock after holding it for a short time period, April/May aside, you were likely realizing a capital gain.

If however you bought and sold a stock after holding it for longer than a year. You were very likely realizing a capital gain and probably a fairly large one

less pronounced with the russell but still likely a capital gain here if selling in 2025.

The bottom line being that capital gains taxes should be strong again this April.

So whats the call then on the non withheld electronic (EFT) and other (paper checks) total receipts for FY 26 which is where the capital gains taxes will be felt?

Last year in early January I projected 1.19T combined across the categories.

The categories finished at 686.3b Electronic and 466.7b Other, so 40b light of my projection. But…… I wrote that piece/projection prior to the LA County fires and their impact on delayed receipts. which looks to be around 30-40b that shifted into FY26 from FY25 so my projection was very close though my split EFT v. Other was a little off since adjusting for the delayed receipts was ~ 715b Electronic, 470b Other. My projection for this year is

Electronic - 810b and Other - 470b

That reflects both non capital gains growth in income tax, strong capital gains tax receipts, shift of ~35b from FY25 to FY26 due to LA County fires as well the impact of lighter receipts due to some of the OBBB tax policy changes.

My projected impact on individual tax refunds is withdrawal growth from ~303b to 370b.

Other impacts to FY26 spending

Tax receipts/refunds are the biggest wildcards for projecting the Jan-Mar and Apr-Jun financing need but there are other receipts and spending changes that are having an impact this year vs. last

When I run a comparison of the current Total Treasury Model projected output for TGA categories on June 30 2026 vs the actual output for the same categories last year, these are the meaningful differences.

Lighter expected corporate income tax receipts more than offset by considerably higher withheld and non-withheld income tax receipts. Plus a significant impact from tariffs.

For withdrawals, the lowered student loan limits are having a meaningful effect on Dept of Education outlays and some of the FEMA and EPA gorge spending early last fiscal year was not repeated.

On the other hand, growth in interest on Treasury securities and growth in social security and medical spending continues to expand.

Putting it all together - Winter QRA projection

Read more

Are similar rises present in the December data and what impact will the 200b in MBS purchases that FNMA/FHLMC are likely to perform in the (nearish?) future have on the delinquent/defaulted FHA loans? This post will both present the December FHA numbers (as always broken down by various portfolio segments so you can understand the internals of whats going on in the portfolio) and explore at a high level the impact of the MBS buys with a more detailed data driven analysis post in the next week or so.

So what happened….

Read more

Turns out I was off by about 5b as the TGA closed at 872.8b on 12/31

The TTM cant perfectly predict the path of the TGA months to a year in advance, but it usually gets pretty close.

So much so that the TTM had outperformed Treasury’s QRA projections (essentially Treasury’s internal models) for the front quarter for both the Spring and Summer QRAs. October thru December results are now in, did the TTM earn the trifecta?

Indeed it did.

Treasury revised their projected Financing Need for Oct-Dec to 531b in the Fall QRA, up 37b from their original 494b projection. Should have stuck with the original since the actual financing need will be right around 486b with Daily Treasury Statement Table II withdrawals clocking in at 1920b for the qtr and deposits at 1434b

In my final QRA projection issued Nov 2nd the TTM projected 507b. Three for three baby!

Aside from the self back patting (but hey, I fully admit/own and retrospect on my mistakes/misses so I think that gives me a little leeway to point out my successes) there is also an interesting angle here. A 486b realized financing need in Q4 25 (even if a bit lower than it would have been due to delayed LA County tax receipts) is down from the 620b realized financing need in Q4 26 (though with an admittedly unfavorable calendar effect that probably added 75b to it)

Perhaps “nothing stops this train”, but it is slowing down some. Will that slowdown continue into 2026? Great question that will be answered later in the piece.

But first Reserve Management Purchases (“RMPs”)

Are they QE John? Meh……… I dont think it matters what they are labeled as. They remove a tiny amount of duration from the available pool of credit risk free assets. They replace what isnt cash (but more or less might as well be), tbills, with actual cash, reserves. Call it what you will. What actually interests me is their impact on Treasury and how much marketable bill issuance Treasury will have to issue.

Mechanically, the RMPs impact Treasury in the same way the rollover of prepaying MBS into treasury bills does. In fact the Feds t-bill purchase schedules doesnt delineate between the t-bill purchases for MBS rollover and RMPs, it just lumps them all together. For the Treasury impact, it is initially zero when the Fed first settles the treasury bill buys from the secondary market. The impact comes about later when the bills that the Fed buys mature, because the Fed will then rollover those bills into new treasury bills issuing on that date (using the same auction add-on rules they have been using for more than a decade). So the Fed add ons will grow for the tbills issuing on the purchased tbills maturity date which means Treasury can reduce their offering size and achieve the same proceeds it needs to keep the TGA at the risk policy level.

So the precise bills the Fed is buying makes a difference in when the Treasury gets “relief” in the future. The Fed splits their buys into two buckets, bills maturing in 1-4 months, and bills maturing in 4-12 months with about 40b in the former and 14b in the latter bucket for the December purchases. In November, I crudely modeled this (well the 15b or so from MBS rollover) figuring the purchases would spread out roughly evenly over the bills the Fed was willing to purchase during an operation and added the projected daily impact of the “Treasury relief” into the TTM so its impact on the TGA levels and thus the cushion above the risk policy level could be gauged. (Note the MBS Tbill Rollover column in the image below from the 12/4 TTM update)

There are two problems with this approach:

1. I was not including the Feds additional tbill holdings for the day (as a result of the RMP/T-Bill Rollover purchases) when calculating things like the SOMA add-on amount projected for the day that I include as part of the model output but more importantly

2. These bill purchases are not proving out to be spread evenly amongst the bill issues the Fed is willing to buy in the operation. For example see the purchases on 12/17

   

   Fed bought 8.165 worth of bills on that day with almost all of them being those that mature on 3/19. I dont believe this is the Fed specifically trying to purchase those bills but rather some market force that caused primary dealers to more aggressively price those bills and the Fed just took the “best” deal.

Regardless, that is certainly not spread out. My problem is, given the admittedly limited purchase data we have so far, I dont really know how to project/model what has happened thus far, moreover Im not really sure its possible, but what I do know, is it is 100% possible to know that the Fed will have 8.065b more bills maturing on 3/19 after the 12/17 purchase settled then they did before, and that matters.

So I have spent the past few weeks updating the model to incorporate these bill purchases at a cusip level. It will be a pain in the ass but I will update the models config after each purchase operation for the bills actually purchased so the TTM when they mature and provide Treasury “relief” (and the projected SOMA add-ons for that day will likewise be correct). As for purchase operations that havent happened yet? Well best I can do right now is spread them over the available issues to buy in the maturity bucket and then update them when the actual purchases become known. Perhaps after a few months a clear and consistent pattern will emerge that allows me to better project the purchase operations that havent happened yet.

What about the purchase amounts beyond the December-January buys? I will have to keep updating the TTM after each purchase schedule comes out around the 9th but for now have the Fed buying ~54b for Jan-Feb, Feb-Mar, and Mar-Apr and after that buying ~22b (15b of which is MBS rollover) through the end of the fiscal year. That roughly matches the Feds guidance on these purchases but will update with actuals as they happen.

Will these RMPs make a material difference? Yeah they will and I think we will see it reflected this month.

Read more

1. Real estate is local. Home values trending up (or down for that matter) where there are very few FHAs or where the FHAs in that locale were mostly originated in the pandemic or earlier and thus already saw a large amount or home price appreciation, wont have much effect on whether an FHA loan goes underwater or not. What matters is the trend in the locales where there are both a large concentration of FHAs and those FHAs have been originated in the past few years (builders…..cough…..builders).

2. Home value appreciation/depreciation is only one piece of the underwater equation. Whether the borrower is paying their mortgage also impacts it. If they are, then home value appreciation being equal, they lower their loan to value a little each month as they pay down principal (although in the first few years of a 30 year amortization its fairly small). If they are not paying though, than they raise their loan to value by a good bit each month since the unpaid interest and escrow portion for that month are now part of the uncollected monies owed on the loan (and no an FHA paid partial claim doesnt change this since that money ultimately still is eventually owed to the FHA albeit at 0% interest). To illustrate, consider a $300,000 mortgage at 6.25%. The interest and escrow amount each month is likely to be around $2,000 each month. So, home values staying steady, a missed payment can add roughly 2/3 of a basis point to the LTV. FHAs saw a large rise in the number of 1 month late folks in November. For any of those folks at or near the 100% LTV line, that missed payment likely took them over it.

So to understand which local real estate markets are most vulnerable to downward home price pressure from underwater FHA borrowers facing a short sale (where the FHA unemotionally sets the acceptable price) or worse with foreclosure. You have to know both where the FHAs are, how they are performing and what home value indexes have done since the loans origination to understand what equity the borrower might have. I present all three.

Read more

Last month I predicted we would see the following in the FHA MBS data for November:

1. A collapse in the number of partial claims

2. A large spike in the delinquency rate

3. An even larger spike (by MoM percentage increase) spike in the serious delinquency rate

Did we see these things?

Read more

The TGA closed on Friday 12/5 with a balance of 870.232b, a bit lower than the 893.388b the Total Treasury Model was projecting in my last update 3 weeks ago but to be expected given the delayed govt shutdown spend (which the TTM did not try to model) leaving the TGA as well as Treasury starting the reductions in the 6w bill 1 week earlier than I had modeled.

With 2 months of the fiscal year under our belt I performed a mini-overhaul of the TTM looking at TGA deposit and withdrawal categories where the models projections were drifting too far away from the realized actuals as of 12/5/2025 and adjusted them to better align with what has happened which then adjusts the future expectations for the category. This post will discuss where the TGA is going over the next month, highlight the larger categories that were impacted and then provide the updated model output through the rest of the fiscal year. Some interesting things are afoot.

TGA Path through early January

Read more

Read more