HUD 223(a)(7) and IRR Loan Modification: How to Reduce Your Rate

The IRR is a recast of the existing HUD mortgage. The mortgage note is amended to a new lower interest rate, but the unpaid principal balance and the remaining term don't change. You're not refinancing the loan — you're modifying it. The only economic levers that move are the interest rate and the resetting of the prepayment penalty.

HUD mortgages have a prepayment penalty that's typically structured as a 10-9-8-7-6-5-4-3-2-1 dropdown: 10% of the unpaid balance in Year 1, 9% in Year 2, declining one percentage point per year, with no penalty after Year 10. Recasting the existing note changes the cash flow going to the GNMA mortgage-backed security investor, so the prepayment penalty is owed at the IRR closing.

In our example, the unpaid balance at the start of Year 6 is $47,290,127. The Year 5 prepayment penalty is 5%, which works out to $2,364,506. The current rate is 5.50% and the market rate is 3.30%.

Both the IRR and the 223(a)(7) are designed to capture the lower rate without the borrower writing a check for the prepayment penalty. The mechanism the IRR uses to absorb that penalty is to artificially elevate the interest rate just enough that the additional interest paid equals the prepayment penalty in present value terms. The MBS investors calculate how much above the market rate the new rate needs to be set so the excess interest income compensates for the penalty. In our example, that math lands the new IRR rate at approximately 3.85% — 55 basis points above the 3.30% market rate, with that 55 bps fully covering the $2,364,506 prepayment penalty. (The exact rate amount changes day to day with Treasury futures, so for live deals reach out to us for current quotes.)

The borrower's rate drops from 5.50% to 3.85%, producing annual debt service savings of approximately $561,698. The closing costs are minimal — title and recording, some borrower legal, and a small Ginnie Mae placement fee, totaling roughly $25,000. With those savings, the borrower recoups the closing costs in less than a month.

The downside is that you're resetting the prepayment penalty. Even though five years of the original 10-year prepay have already burned off, the IRR generally restarts the penalty back to a fresh schedule. If the plan is to dispose of the asset or do a cash-out refinance in the next few years, the IRR may not pencil — the new prepay will eat into the savings. If rates have dropped enough that the savings outpace the prepay reset, it can still work, and we can run that math quickly.

Two mechanical points worth noting on the IRR. First, because it's a recast of the existing mortgage, it can only be done with the existing HUD lender. Second, there are minimal third-party reports — only a cheap update to the HUD Environmental forms. The full transaction generally closes in around 60 days.

The 223(a)(7) has the same goal as the IRR — capture the lower market rate without requiring borrower cash for the prepayment penalty — but the math allows for a lower interest rate because the 223(a)(7) is a formal refinance rather than a recast. That distinction matters because of how the prepayment penalty gets absorbed.

Under the 223(a)(7), the new loan amount can be increased up to the original loan balance. In our example, the original loan was $50,000,000 and the current unpaid balance is $47,290,127, leaving $2,709,873 of headroom in the new mortgage. That headroom can be used to fund financing fees and a portion of the prepayment penalty as part of the new loan rather than absorbing all of it into rate.

HUD prioritizes financing fees first in that headroom. Financing fees on a 223(a)(7) typically run around 1% of the loan amount — lender fees, HUD application fee, HUD initial MIP, third-party reports, title and recording, GNMA placement fee, and legal. In our example, that's approximately $500,000 in financing fees, leaving roughly $2,209,873 of remaining headroom to apply against the $2,364,506 prepayment penalty. The shortfall — only $154,633 — is what has to be absorbed into rate.

Because we're only covering $154,633 in rate (versus the IRR's $2,364,506), the rate needed above the 3.30% market rate is far smaller. The new 223(a)(7) rate lands at approximately 3.34% — just 4 basis points above market. Compared to the IRR's 3.85% and the original 5.50%, the 223(a)(7) produces annual debt service savings of approximately $791,911, or roughly $230,000 per year more than the IRR.

The 223(a)(7) also lets you extend the term back to the original loan term, provided you're not more than 12 years past the original loan's first payment date. In our example, the borrower closed five years ago on a 35-year amortization, so the new 223(a)(7) gets a fresh 35-year term. A longer amortization spreads the same loan amount over more periods, lowering monthly P&I — another reason the 223(a)(7) produces a lower payment than an IRR for an equivalent loan size.

As with the IRR, you have to start a new prepayment penalty schedule at closing. There's no way around that — it's the price of capturing the lower rate.

The 223(a)(7) has real downsides relative to the IRR that the borrower needs to weigh. Because it's a refinance and not a recast, legal and title and recording fees are larger, and HUD requires a Property Capital Needs Assessment (PCNA). The PCNA can surface deferred maintenance, items in the existing CapEx schedule that are coming due sooner than originally projected, or other physical conditions that may require a step-up in the replacement reserve escrow. Some borrowers prefer to address those items on their own rather than re-invite HUD into the conversation. There are also more HUD fees because HUD is actually re-reviewing the application and reports — and that review adds time. A 223(a)(7) generally takes a couple of months longer to close than an IRR.

Finally, the 223(a)(7) increases the borrower's debt liability on the balance sheet. The IRR keeps the unpaid principal balance the same. The 223(a)(7) steps it up to absorb financing fees and (typically) part of the prepayment penalty. That extra debt becomes more relevant at exit — whenever the asset is sold or refinanced, the higher payoff has to be funded out of sale proceeds or new debt.

Here's the head-to-head comparison using the example deal — the $50,000,000 mortgage that closed five years ago at 5.50% with a current unpaid balance of $47,290,127, a 5% prepayment penalty of $2,364,506, and a 3.30% market rate.

Looking only at annual debt service savings, the 223(a)(7) wins by a wide margin. Looking at speed, simplicity, and balance sheet cleanliness, the IRR wins. Which is the right call depends on the borrower's plans for the asset — and there's a third option worth running the math on before deciding.

When we have the IRR vs. 223(a)(7) conversation with borrowers, we often end up recommending neither — at least not in their pure form. The structure that frequently produces the best NPV over a typical hold is what we coined as the "Hybrid 223(a)(7)."

The mechanic is straightforward. Recall that under a standard 223(a)(7), the loan amount can step up to the original balance to absorb both financing fees and a portion of the prepayment penalty, with the small remainder absorbed into rate. In a Hybrid 223(a)(7), we cap the loan amount at just enough to cover the financing fees — and absorb the entire prepayment penalty into rate, the way an IRR does.

This is not how most underwriters or originators size a 223(a)(7). The standard underwriting workflow runs three loan amount tests — the original loan amount, the cost to refinance, and the debt service constraint at 90% of NOI — and takes the lowest. In the rate environment that drives 223(a)(7) volume in the first place, the original loan amount or cost to refinance is almost always lower than the debt service constraint, so that becomes the binding test and the loan is sized to it. Intentionally walking the loan amount down from there to the unpaid balance plus financing fees only — leaving headroom on the table to push the prepayment penalty into rate instead — is not something the standard workflow produces, because the standard workflow takes the lowest of the tests and stops. Sizing the Hybrid requires deciding to override that output, and most originators don't think to do it.

The result is a 223(a)(7) with three characteristics:

A higher rate than a fully maximized 223(a)(7). Because we're absorbing the full prepayment penalty into rate rather than splitting it between rate and loan amount, the new rate sits above what a maximized 223(a)(7) would produce.

The original loan term, not the remaining term. Like a maximized 223(a)(7), the Hybrid extends the amortization back to the original loan's term — 35 years in our example, versus the IRR's 30 remaining years. That longer term lowers the monthly P&I payment for any given loan amount and rate.

A debt liability close to the IRR's, not the maximized 223(a)(7)'s. The loan amount is roughly equal to the current unpaid balance plus financing fees only — a much smaller step-up than a maximized 223(a)(7), which adds the full prepayment penalty's worth of new debt on top.

When we model this against a determined hold period — where the asset is either sold or refinanced into something else at the end of the period — the NPV ranking is usually Hybrid 223(a)(7) > IRR > maximized 223(a)(7). The reason is mechanical. The maximized 223(a)(7) produces the highest annual debt service savings, but at exit the borrower has to fund a meaningfully higher payoff balance — and over a typical hold, that extra debt liability outweighs the additional rate savings. The IRR keeps the balance low and gives up some of the rate benefit. The Hybrid threads the needle: most of the rate benefit, the longer term, and a debt liability close to the IRR's.

The Hybrid isn't always the answer. In a deep rate-drop environment where the cash savings dominate, the maximized 223(a)(7) can still win on NPV. If the borrower expects to hold the asset effectively forever and never refinance or sell, the lower rate of the maximized 223(a)(7) compounds. And in scenarios where speed matters — capturing a brief rate dip before it disappears — the IRR's faster timeline can outweigh the rate disadvantage.

The point is that the IRR vs. 223(a)(7) framing is incomplete. There are three structures, not two, and the Hybrid is the right answer more often than most borrowers expect. We can run all three side by side on your specific deal in a few hours.

The right structure depends on three things: how long you plan to hold the asset, how much rates have dropped, and how concerned you are about debt liability at exit. A few common scenarios:

Capturing a quick rate dip. When market rates open a window that may not stay open long, the IRR's speed advantage matters. A 223(a)(7) takes around four months to close — long enough that rates can move materially during the process. An IRR can close in roughly half that time and locks the rate sooner. If the rate move is meaningful but not enormous, and you want to capture it before it disappears, the IRR is often the right tool.

Long-term hold with rates well below the existing rate. When the market rate is materially below the existing rate (200 bps or more in our example) and the plan is to hold the asset for the long run, a maximized 223(a)(7) captures the most savings over time. The PCNA risk is real but manageable, and the additional debt liability becomes less relevant the further out the eventual disposition is.

Anticipated sale or cash-out refinance in the medium term. When the holding period is meaningfully shorter than 10 years — say a 5-to-7 year exit horizon — the Hybrid 223(a)(7) usually wins. You get most of the rate benefit and the longer term, without taking on extra debt that has to be paid off at exit.

None of these are hard rules. They're the patterns we see most often. Every deal has specifics that shift the math, and we model all three structures before recommending one.

Can I do a 223(a)(7) or IRR with a different lender than my current one?
For an IRR, no — it's a modification of the existing note, so it has to be done by the lender currently servicing your HUD mortgage. For a 223(a)(7), yes — it's a refinance, and any HUD-approved MAP lender can originate it. If you're not happy with your current lender's service or pricing, the 223(a)(7) is your route to a different shop.

Can I take cash out with a 223(a)(7) or IRR?
No. Both are explicitly cash-flow refinances. Their entire purpose is to reduce debt service and increase the borrower's cash flow — not to monetize equity. If you want to pull equity out of a HUD-financed asset, you would need a regular 223(f) cash-out refinance, which is a different program with different mechanics and a different sizing analysis.

How long does a 223(a)(7) take to close versus an IRR?
An IRR generally closes in around 60 days. A 223(a)(7) generally takes around four months from engagement to closing — driven mostly by HUD's review timeline and the PCNA delivery. The Hybrid 223(a)(7) follows the same timeline as a standard 223(a)(7) — it's structured the same way, just with a smaller loan amount.

What happens to my existing prepayment penalty when I do an IRR or 223(a)(7)?
The existing prepayment penalty is paid off at closing. In an IRR, the entire penalty is absorbed into a slightly elevated interest rate. In a 223(a)(7), part of the penalty is funded by increasing the loan amount up to the original loan balance, and any remainder is absorbed into rate. The borrower doesn't write a check for the penalty in either case. Both structures then reset the prepayment penalty schedule on the new loan, generally back to a fresh 10-year 10-9-8-7-6-5-4-3-2-1 dropdown.

Can I do a 223(a)(7) on a 221(d)(4) loan, or only on a 223(f)?
Yes — the 223(a)(7) is available on any existing HUD-insured multifamily mortgage, including 221(d)(4) construction loans that have converted to permanent or even an already done 223(a)(7). The mechanics are the same.