Real estate debt and CMBS, at coverage banker depth

The capital stack, top to bottom

Every real estate financing conversation starts from the same picture: one asset, one stream of net operating income, and a stack of claims on that stream ordered by who gets paid first and who takes the first dollar of loss. Draw the stack, explain why each layer exists, and you can handle most of the debt questions a real estate, gaming and lodging group will ask.

Senior mortgage. A first lien recorded against the real property itself. It is the cheapest money in the stack because the remedy is the cleanest: if the borrower stops paying, the lender forecloses and owns the building. Mortgage lenders size conservatively and expect to be the only lien on the asset, which is why the layers above them had to be invented.

Mezzanine debt. Mezz sits behind the mortgage in priority, but it is not secured by the property. The senior loan documents forbid a second lien on the real estate, so the mezz lender lends one level up the ownership chain, to the parent that owns the equity interests in the property-owning borrower, and takes a pledge of those interests as collateral. On default it forecloses on the pledged membership interests under the Uniform Commercial Code, a matter of weeks rather than the months a real property foreclosure takes. What it ends up owning is the entity that owns the building, with the mortgage still in place and still due. That is the whole point and also the whole risk: to take the keys, the mezz lender has to keep the senior current out of its own pocket. An intercreditor agreement sets out its cure rights, its right to purchase the senior loan, and the conditions on foreclosure.

Preferred equity. Economically this looks like mezz, with a stated return that accrues and priority ahead of common. Legally it lives inside the ownership entity as equity rather than debt, so the remedy is usually a change of control (the preferred holder removes the managing member) rather than a foreclosure. Sponsors reach for pref when the mortgage documents prohibit even a pledge of equity interests, which closes the door on mezz.

Common equity. Sponsor capital plus limited partner capital. Last money in the loss stack, all of the residual upside, and the promote that pays the sponsor for outperformance.

The trap here is small and catches a lot of candidates. Do not call mezzanine "a second mortgage." Say the equity pledge out loud. It is the sentence that tells the interviewer you have actually seen a real estate capital structure.

Property level secured debt versus corporate unsecured debt

A private sponsor with eight buildings usually has eight separate mortgages, each on its own asset in its own borrowing entity. A large investment grade REIT with two hundred buildings does not run its balance sheet that way. It migrates toward corporate unsecured debt: bonds, a revolving credit facility, and unsecured term loans in between.

The main reason is flexibility, not pricing. Selling a mortgaged building means paying off, defeasing, or getting the lender's consent to an assumption, and refinancing one asset means renegotiating with one lender who knows you are on a clock. A REIT is in the business of recycling capital, and every recorded mortgage is a veto point on that. Unsecured debt has no lien and therefore no consent right over any individual asset, so the company can transact freely as long as it stays inside its covenant package. (Those covenants work like any corporate credit agreement, and the leveraged finance terms guide covers the mechanics properly.)

The second reason is access. REITs are habitual capital markets issuers because they cannot retain much cash: the distribution requirement pushes taxable income out every year, so growth is funded externally rather than from retained earnings, which REIT structures and tax rules covers in depth. A company that has to come to market repeatedly cares intensely about keeping that market open.

The concept tying both together is unencumbered asset coverage. An unsecured bondholder has no lien on anything. What protects it is the pool of assets not pledged to a mortgage lender, since those are what unsecured claims reach in a restructuring. So REIT bond indentures carry an unencumbered asset test, customarily requiring unencumbered assets worth at least 150 percent of unsecured debt. That value is usually estimated by capitalizing the NOI of the unencumbered properties, so the test moves with cap rates as well as cash flow, which connects to how real estate companies are valued.

The good answer to give is that mortgaging a building costs the REIT twice, once for the lien and once for pulling the asset out of the pool supporting the unsecured rating. Bankers read a rising share of secured debt as a stress signal, not a financing choice.

The three metrics every real estate lender underwrites

Every real estate lender, on every asset class, sizes a loan against three constraints and lends the smallest answer.

Debt yield is the one worth understanding properly, because it is the only one with no appraisal, no interest rate, and no amortization schedule in it. Nothing about how the loan is structured can move it. A $100M loan on a property producing $10M of NOI is a 10 percent debt yield whether the loan is fixed or floating, amortizing or interest-only, and regardless of what the appraiser thought the building was worth. That is why lenders treat it as the honest measure, and why the debt yield floor rather than the LTV cap is often what actually sizes the loan. NOI carries a lot of definitional weight in these tests, and cap rates and net operating income is where the line items get pinned down.

Recourse, non-recourse, and the special purpose entity

Most commercial mortgage debt in the United States is non-recourse. The lender's remedy runs to the property and the rents, not to the sponsor's other assets, which is what makes it possible to raise third party equity into a single deal.

Non-recourse is never absolute, though. Every such loan carries bad-boy carve-outs, backed by a guaranty from a creditworthy parent or the sponsor personally, plus a separate environmental indemnity. They come in two tiers. The lighter tier makes the guarantor liable for actual losses from specific bad behavior: fraud, misapplication of rents or insurance proceeds, physical waste, or failure to pay taxes when there was cash to pay them. The heavier tier makes the entire loan fully recourse and is reserved for a short list, most commonly a voluntary bankruptcy filing, an unpermitted transfer of ownership interests, and an unpermitted subordinate lien. That springing tier is not there to collect. It is there to make bankruptcy so expensive for the sponsor personally that the sponsor negotiates instead.

The special purpose entity requirement is the structural half of the same idea. The borrower must be a newly formed entity that owns one asset, has no other business and no other material debt, keeps its own books and bank accounts, does not commingle funds, and observes separateness formalities, often with an independent director whose consent is required before it can file for bankruptcy. Ring-fencing keeps a problem at one property from infecting the sponsor's other assets, and lets the lender underwrite one building rather than the credit of a private sponsor it cannot see through. Breaching the SPE covenants is itself a bad-boy trigger, which is what makes the structure hold together.

Construction debt, transitional debt, and agency multifamily

A development has no NOI at closing, so there is nothing to size a DSCR against. Construction lenders underwrite cost instead of value, size to loan-to-cost rather than loan-to-value, and require the sponsor's equity to fund first or alongside. The loan funds in draws against verified progress, with an inspecting engineer signing off and retainage held back. Because the asset produces no cash, an interest reserve is sized into the budget and the loan effectively pays its own interest until stabilization. On top sits a completion guaranty obligating the sponsor to finish the building lien-free, plus often a carry guaranty for shortfalls.

A finished but unleased building is not financeable as a permanent loan, so a bridge or transitional loan carries it through lease-up: floating rate, interest-only, a shorter term with extension options, and future funding for tenant improvements, leasing commissions and capital work. Once cash flow supports a coverage test, it refinances into a permanent fixed rate loan or a securitized execution. Construction, then bridge, then permanent: naming the three stages handles most development financing questions.

Multifamily deserves a separate mention. Agency financing through the government-sponsored enterprises gives apartments a debt market no other property type has. Loans are originated through networks of approved lenders under standardized programs, then purchased and securitized by the agencies, so terms are uniform, tenors are longer, leverage runs higher than conventional lenders offer, and the market stays open in parts of the cycle when balance sheet lenders pull back. The coverage takeaway is that multifamily is more reliably financeable than anything else in real estate, which supports both transaction volume and pricing.

CMBS at coverage depth

In a commercial mortgage backed securities execution, individual mortgage loans on commercial properties are pooled into a trust, and the trust issues bonds sold to investors. Two shapes dominate: conduit deals aggregating many loans from many borrowers, and single-asset single-borrower deals built around one large loan. Selling the risk to bond investors rather than holding it on a balance sheet lowers the borrower's all-in cost, which is why sponsors use it.

You do not need the tranche mechanics for a coverage interview. You need the consequence: the borrower trades flexibility for cost, and the flexibility loss is severe.

Once the loan is in a trust, there is no lender to call. Routine administration goes to a master servicer that collects payments and administers reserves. On default, or imminent default, the loan transfers to a special servicer whose discretion is bounded by the pooling and servicing agreement. It tests modifications against a net present value standard for the benefit of bondholders, answers to a directing certificateholder with its own economic interest, and moves at the pace of a documented process rather than a relationship. Modifications happen, but slowly and on the servicer's terms.

Prepayment is the other half. Securitized loans are typically locked out for a period, after which the borrower gets out through defeasance (substituting government securities that replicate the loan's payment stream so the trust keeps its cash flow) or a yield maintenance payment compensating the lender for lost interest. Neither is a simple payoff at par, both carry cost and process, and defeasance cost swings with the rate environment.

That is the piece that shows up in your work. In a sale process or a take-private, the buyer usually wants the existing debt either gone or assumed. Gone means defeasance or yield maintenance, a real cash cost out of equity value. Assumed means servicer approval, rating agency confirmation, an assumption fee, and a timeline on the servicer's clock rather than the deal's. Either way the loan becomes a diligence item, a timing risk and a price adjustment, identified in week one and modeled rather than discovered at signing. The same friction shapes REIT M&A and take-privates, where a target's secured debt stack can drive the structure of the whole transaction. The bond mechanics are a fixed income topic; the inflexibility is the client conversation.

Rate exposure and ground leases

Permanent loans are usually fixed rate, and construction and bridge loans float over a benchmark, which follows from the business plan: a stabilized asset wants payment certainty, a transitional asset wants prepayment flexibility while the sponsor executes. Floating rate lenders almost always require the borrower to buy an interest rate cap, with the strike and term written into the loan documents. It is a condition to closing, it has to be maintained, and it has to be replaced when the borrower exercises an extension option. Replacement cost is set by market conditions at the time and can be far larger than the original, a common surprise for sponsors relying on extensions. Corporate term loans manage the same exposure with swaps, which carry breakage cost.

Ground leases change both financeability and value. The fee owner keeps the land, and the building owner holds a leasehold interest for a long term. Lenders will finance a leasehold, but they want the remaining term to run well beyond loan maturity, they want leasehold mortgagee protections in the lease (notice and cure rights, and the right to a new lease if the ground lease is rejected in the fee owner's bankruptcy), and they underwrite the rent reset because it can consume a large share of NOI. A leasehold with a shortening term or weak protections gets less debt and trades at a wider cap rate than the equivalent fee simple asset, which often explains an apparently mispriced comparable.

Practice question

Lenders already look at LTV and DSCR. Why do they also look at debt yield, and which one actually sizes the loan?

Debt yield is NOI divided by the loan amount, so a $100M loan on a building producing $10M of NOI is a 10 percent debt yield. What makes it useful is what is not in it. There is no appraised value, no interest rate, and no amortization schedule, so nothing about how you structure the loan can move it. LTV depends on an appraisal, and appraisals are built off cap rates, so falling cap rates let the same building support a larger loan. It looks best exactly when the market is most aggressive. DSCR depends on the rate and the amortization, so interest-only floating rate debt can show comfortable coverage that falls apart on a reset. Debt yield strips both out and answers one question: if I foreclose tomorrow, what unlevered yield do I own at my basis?

In practice a lender runs all three, gets three loan amounts, and lends the smallest. When values are high and rates are low, the debt yield floor usually caps proceeds, because LTV and DSCR both look generous on the same cash flow.

What the interviewer is listening for: That you can compute all three without hesitating, and that you know why debt yield is the honest one, which is the absence of any loan or valuation input. The best answers size to the minimum of the three rather than picking a favorite metric. Naming which constraint binds in which environment separates a candidate who memorized formulas from one who has thought about loan sizing.