What the first-half data shows
Average loan-to-value ratios on US commercial real estate loans rose to 65.9% in the first half of 2026, up 170 basis points from 64.2% a year earlier, according to MSCI data reported by GlobeSt and CRE Daily on October 1, 2026.
The write-ups are GlobeSt's and CRE Daily's. Bisnow carried the same 64.2% prior-year comparison on October 5. Every MSCI figure in this piece comes from that reporting.
Underneath the average, banks came back. They accounted for 39% of commercial property lending in the first half, up from 34% a year earlier, which puts their share back in line with its 10-year average. National banks moved fastest. Their lending volume rose 75% year over year, and their share rose 3 points to 14%.
Where the leverage is coming from
The top of the leverage range belongs to investor-driven lenders, the MSCI category that includes debt funds. Their share rose to 16% from 13%, and their average LTV was 69.5%, the highest of any lender group MSCI tracks.
Agencies went the other way on apartments. Per the same MSCI data, government agencies accounted for 38% of apartment originations, down 12 points from a year earlier and well below the roughly 50% they averaged over the past decade. Agency LTV rose to 63% from 61%. They are lending a little deeper on the deals they win and winning a smaller share of the market.
Insurance companies posted the biggest LTV jump of any group, up 250 basis points to 62.7%. Their share slipped to 9%.
Read together, banks and debt funds are winning apartment business that used to default to an agency execution. For a sponsor, the lender across the table on the next financing is more likely to be a bank or a fund than it was a year ago.
Two datasets, two answers
CBRE's data points the other way. In its Q2 2026 U.S. Capital Markets Report, multifamily LTVs eased to 63.3% from 65.8% a year earlier, and commercial LTVs eased to 59.6% from 60.8%. CBRE's LTV figures cover fixed-rate permanent loans only.
The two series count different loans. MSCI counts originations across lender types, bridge and debt-fund paper included. CBRE's LTVs are permanent fixed-rate debt. A market in which debt funds average 69.5% while permanent lenders hold multifamily leverage flat or lower would produce exactly this split.
MSCI's Jim Costello has also cautioned that the data does not conclusively show lenders taking materially more risk. In his read, part of the rise may reflect a shifting mix of loan sizes and property types, not looser terms on any given loan.
Our reading is narrower than the headline. Leverage is loosening at origination, in the bridge and balance-sheet channels. It is not clearly loosening on permanent takeouts. A sponsor's exit sits on the second of those.
What it means at $1M to $5M
Most of the national-bank surge is not happening at this loan size. Regional and local banks handled 60% of loans of $10M or less in the first half, and their overall share rose to 21% from 18%. Their average loan was roughly $6.4M.
That fits what the bank data has been saying all year. In August we wrote that the easing in the Fed's loan officer survey was concentrated at the largest banks, and in July we looked at how much CRE room New Jersey's banks actually have. The biggest lenders are growing fastest. The small-loan market still runs through regional and local banks, and debt funds are now competing for the same files from the other side.
That competition is the sponsor's leverage on bridge structure: proceeds, holdbacks, extension terms. A file with three interested lenders negotiates differently from a file with one. Use it on the loan you are signing.
Worked example: when coverage sets the exit
A permanent loan is sized to whichever test binds first, loan-to-value or coverage. The LTV on a term sheet is a ceiling. Whether you reach it depends on the loan constant.
A loan constant is the annual debt service divided by the loan amount, so it captures both interest and principal and depends on the rate and the amortization period.
The example's terms are 65.9% LTV, the first-half market average, 1.25x coverage and 30-year amortization. They are assumptions for this example, not any program's current parameters. One input drives every figure: a purchase cap rate of 6.00%.
| Step | Figure |
|---|---|
| Stabilized value | $4,000,000 |
| NOI at a 6.00% cap | $240,000 |
| LTV-sized loan at 65.9% | $2,636,000 |
| Maximum debt service at 1.25x | $240,000 ÷ 1.25 = $192,000 |
| Maximum loan constant for the full $2,636,000 | $192,000 ÷ $2,636,000 = 7.284% |
| Rate that produces a 7.284% constant on 30-year amortization | about 6.11% |
The takeout reaches the full 65.9% only at a rate of about 6.11% or lower. Above that, coverage sets the loan.
The same arithmetic generalizes. Coverage, not LTV, sizes the loan whenever the purchase cap rate is below 1.25 times the LTV times the loan constant.
| Cap rate | Max loan constant at 65.9% LTV, 1.25x | Break-even takeout rate, 30-yr amortization |
|---|---|---|
| 5.50% | 6.677% | 5.31% |
| 6.00% | 7.284% | 6.11% |
| 6.50% | 7.891% | 6.88% |
| 7.00% | 8.498% | 7.63% |
Break-even rates are rounded down, so each "at or below" reading stays true.
On a multifamily property bought at a 6.00% cap rate, a 65.9% loan-to-value takeout sized at 1.25x coverage on 30-year amortization reaches full leverage only at a rate of about 6.11% or lower.
Run your own deal against the table. Find the cap rate you bought at, then put the rate on your takeout term sheet next to the break-even in that row. If the term-sheet rate is higher, the coverage test sizes your exit and the headline LTV is not the number you will get. The lower the cap rate you paid, the lower the rate you need. We walked the same coverage cap through a full takeout file in the Conventional Small takeout piece.
Using the competition without overreaching
Use the competition on the bridge. Proceeds, holdback size and extension terms are where a lender who wants the file gives ground, and the first-half data says banks and debt funds are pressing for share.
Underwrite the exit to coverage from day one. Size the business plan to the NOI that clears the break-even at a takeout rate you can defend. If the plan only works at headline LTV on exit, it is a bet on the rate, and the bridge lender is not taking that bet with you.
For a bridge already in place and maturing, the arithmetic runs from the balance. The coverage-sized takeout sets how much of it refinances, and the rest is a check. The ratio that sizes that check is debt yield, which we laid out in the cash-in refinance piece.
Bridge proceeds above what coverage will support at the exit are not permanent leverage. They are a short-dated position that NOI has to grow into before maturity, at a takeout rate nobody has quoted yet. The higher the bridge LTV you negotiate, the more of the business plan has to land before the coverage test lets you out.
FAQ
What was the average commercial real estate loan-to-value ratio in the first half of 2026? 65.9%, up 170 basis points from 64.2% a year earlier, according to MSCI data reported by GlobeSt and CRE Daily on October 1, 2026.
At what rate does DSCR start limiting a multifamily refinance? It depends on the cap rate. At a 6.00% cap rate, a 65.9% LTV takeout sized at 1.25x coverage on 30-year amortization reaches full leverage only at a rate of about 6.11% or lower. Above that, coverage sets the loan amount.
Are agencies losing multifamily market share? Yes. Government agencies accounted for 38% of apartment originations in the first half of 2026, down 12 points from a year earlier and below their roughly 50% average over the past decade, per MSCI data reported by CRE Daily.
MSCI figures are MSCI data as reported by GlobeSt and CRE Daily on October 1, 2026; the primary MSCI report is not quoted here. The 64.2% first-half 2025 comparison was also reported by Bisnow on October 5, 2026. CBRE figures are from CBRE's Q2 2026 U.S. Capital Markets Report and cover fixed-rate permanent loans only. The worked example's 65.9% LTV, 1.25x coverage and 30-year amortization are assumptions for the example, not any program's terms. Loan constants are rounded to three decimals and break-even rates are rounded down to two. Every deal sizes to its own underwriting on the day it is quoted.