Yield Maintenance vs. HUD Step-Down: Which Is Cheaper?

Most HUD loans carry some version of the standard step-down. Roughly 90% of the deals we see close with the full 10-9-8-7-6-5-4-3-2-1, where the prepayment penalty is 10% of the outstanding balance in year 1, 9% in year 2, and steps down one percentage point each year until it hits 1% in year 10. After the tenth anniversary of the loan, there is no prepayment penalty for the remainder of the term. The percentage applies to the unpaid principal balance at the time of payoff, not the original loan amount, so the dollar amount of the penalty decreases over time both because the rate is stepping down and because the balance is amortizing.

Alternative call protection structures are allowed and commonly requested. Instead of the full 10-9-8-7-6-5-4-3-2-1, a borrower can negotiate a Lockout-Lockout-Lockout-7-6-5-4-3-2-1, a Lockout-Lockout-4-3-2-1, a 7-8-6-5-4-3-2-1, or any number of other variations tailored to a specific business plan. The general rule is that the more favorable the call protection is to the borrower, the higher the interest rate will be. Whether it pencils depends on how likely a payoff event is within the prepayment window and what the rate add actually is at current pricing. Reach out and we'll run the side-by-side on current coupons.

The most important feature of the step-down for cash flow modeling is that the penalty in each month is fixed at origination. It doesn't move with treasuries, it doesn't move with credit spreads, it doesn't move with anything. If you want to know what the penalty will be in month 47 of your loan, you look at the scheduled balance in month 47 and multiply by 7%. Done.

Most agency multifamily loans use some version of yield maintenance. Fannie Mae uses a yield maintenance formula directly. Freddie Mac typically uses defeasance, which substitutes a portfolio of treasuries matching the loan's remaining cash flows in place of the mortgage — structurally different, but economically very similar to yield maintenance because the cost of replacing the lender's income stream tracks the same rate math. For the rest of this paper we'll refer to both as "yield maintenance" and use the Fannie Mae calculation.

The FNMA yield maintenance formula is:

Where UPB is the unpaid principal balance at the payoff date, Reinvestment Yield is interpolated from the on-the-run treasury curve between the two maturities bracketing the remaining yield maintenance period, and PVA is the present value annuity factor over the remaining years discounted at the reinvestment yield. The result is floored at 1% of the UPB — meaning even if rising rates would produce a zero or negative YM number, the borrower still pays 1%.

Rather than fighting through the symbolic math, walk through a concrete example. Say you have the same $25M loan at 5.30% and you want to pay it off on the loan's fifth anniversary. Your balance at that point is roughly $23.6M. The loan has a 9.5-year yield maintenance period, so there are roughly 4.5 years remaining on the YM window at that payoff. Interpolating the treasury curve between the 3-year treasury and 5-year treasury points gives you the reinvestment yield — call it 3.89% under the assumed curve. The spread between the current interest rate and the reinvestment yield is 5.30% − 3.89% = 1.41%. Apply the 4.5-year annuity factor at 3.89%, multiply by $23.6M × 1.41%, and you get roughly $1.38M of yield maintenance. That's the mechanical calculation Fannie does. The present value factor is doing the heavy lifting — it's compensating the lender for the shortfall in yield over the remaining years, discounted back to today.

The percentage points attached to the HUD step-down — 10%, 9%, 8% — sound high, and we think that's where the bad reputation comes from. Yield maintenance doesn't quote a percentage, so it feels softer by comparison. But once you convert both to dollars against the same loan, the comparison looks different.

The assumptions below match the reference model. Every HUD dollar and YM dollar in the chart comes straight out of that model with no smoothing.

Flat Treasuries — The Apples-to-Apples Case

The most honest comparison assumes treasuries don't move over the life of the loan. Predicting where treasuries go is not something we or anyone else should be staking borrower decisions on — if we knew, we'd be trading treasury futures. So the starting point is to run the comparison under the assumption that the treasury curve at the loan origination is the treasury curve at every payoff date.

The shape of both curves is what you'd expect, but the relationship between them is the interesting part. HUD is more expensive in the early years — the 10% penalty on ~$25M is $2.50M in year 1 versus roughly $1.90M–$1.95M of YM. HUD stays more expensive than YM through all of year 4. That gap is there, but it's also largely academic — owners rarely refinance or sell a stabilized asset in the first four years of any fresh permanent loan, HUD or agency or life company. The rate was just locked, the deal was just closed, and the business plan is still running.

Starting in year 5, the picture flips. From month 49 (the start of year 5) through roughly month 80, HUD is consistently cheaper than YM — meaningfully so, usually by $100K–$175K depending on where in the year you measure. Inside each of those middle years you can see a sawtooth on the gold line: the HUD penalty holds at a fixed percentage (7%, then 6%, then 5%) of a slowly-amortizing balance, then steps down at the next anniversary. YM declines smoothly because the annuity factor shrinks continuously. The result is that YM catches up at the very end of each year, right before the next HUD step-down — specifically the last few months of years 5 through 7. In practice, if a borrower is within a few months of an anniversary, they almost always wait for the next drop rather than pay the higher percentage.

Year 8 is roughly a coin flip — HUD wins the first half of the year, YM wins the second half as the remaining YM period gets short enough that the rate-differential × annuity-factor calculation collapses below the 1% floor. From that point on, YM is effectively locked at 1% of UPB, and that's what makes it cheap relative to the HUD step-down that's still stepping down through 2% and 1%. Through years 9 and 10, YM is cheaper. Said differently, once you're in years 9 or 10, many borrowers simply elect to wait out the end of the prepayment period rather than pay either penalty — the window is short enough that a delayed refi or sale is usually the cleaner outcome.

What if Treasuries Move?

Flat treasuries is a clean analytical baseline, however, it is not the real world. So it's worth looking at what happens to the YM penalty if the reinvestment yield moves after rate lock. The HUD line doesn't change, no matter what treasuries do. The YM line moves meaningfully. Click the two buttons below to toggle between treasuries moving up 50 basis points or down 50 basis points.

If treasuries drop 50 bps after rate lock, the reinvestment yield in the YM formula drops, the rate differential widens, and yield maintenance gets materially worse. HUD now beats YM across every month of years 1 through 8, typically by $400K–$600K. Years 9 and 10 still tilt toward YM, but by smaller margins than under the flat case. This is the scenario a borrower is exposed to if they lock a loan and then the economy softens — the call protection tightens on the agency side just when a sponsor would most want the flexibility to refinance.

If treasuries rise 50 bps after rate lock, the reinvestment yield rises, the rate differential narrows, and YM bottoms out at the 1% of UPB floor for most of the loan term. In this scenario YM is cheaper than the HUD step-down in every month of every year — roughly $1.5M cheaper at year 1, $500K cheaper at year 5, and still meaningfully cheaper through year 9. In a material rising-rate environment, yield maintenance is the cheaper form of call protection. This is precisely the window — 2022-2025, broadly — during which the sentiment that "HUD prepayment is worse" became entrenched.

Price aside, the HUD step-down has a second advantage that doesn't show up in any single-scenario chart: certainty. When you're building the cash flow model for a deal, you know in month 1 exactly what the prepayment penalty will be in months 13, 36, 60, and 84. It's the outstanding balance times a known percentage. That number goes straight into your IRR sensitivities for a refinance-at-year-5 case, a sale-at-year-7 case, or whatever disposition scenarios you're running.

Yield maintenance can't be modeled that precisely because it depends on what treasuries do between now and the payoff date. Sophisticated sponsors handle this by running YM scenarios at plus and minus 50 or 100 bps — which is useful, but it's still a range rather than a number. The step-down gives you a number. For sponsors whose capital stack models depend on confident IRR projections to close with investors, that certainty has real value independent of expected cost.

This is also why the step-down is helpful in asset-sale underwriting on the buyer side. A buyer assuming the existing HUD debt knows exactly what call protection is in place at any future point. A buyer assuming Fannie paper with YM needs to model a rate path.

One more thing worth flagging on the topic of HUD prepayment, because it changes the practical calculus of being in HUD paper when rates fall. HUD has two streamlined refinance products — the HUD 223(a)(7) streamlined refinance and the HUD Interest Rate Reduction (IRR) loan modification — that are specifically designed to let an existing HUD borrower capture a rate drop without writing a check for the prepayment penalty. Both allow the existing prepayment penalty to be rolled into the new loan rather than paid in cash at closing. The 223(a)(7) is a full refinance into a new HUD loan; the IRR is a modification of the existing loan's rate without re-originating the note (read about the 223(a)(7) and IRR here).

That combination — a step-down penalty that is often the cheaper form of call protection, plus two streamlined HUD-to-HUD products that let you restructure without coming up with cash for the penalty — is a better deal than the conventional wisdom suggests.