When a mortgage broker's rate advantage actually justifies the fee, and when it doesn't

What did the broker actually do on the mixed use file that you couldn't? I don't mean that as a dig, I'm trying to work out where the labor is. Was it knowing that one lender existed, or was it framing the vacancy in a way underwriting would accept?

The "who's funding this week" line is the whole product as far as I can tell. Everything else is a website. A lender list you built in March is a historical document, not a sourcing tool.

My guy funds off his own balance sheet and I've done four with him. No broker in sight and the terms haven't moved. But I also haven't brought him anything he'd have to think about, which is probably the actual point ledger's making.

So the rule is roughly this. If the deal is shaped like the last one, go direct. If a lender has to think, pay the point. Is it really that clean, or does that framing flatten something in the odd files?

Worth flagging that who can even charge you that point depends on where the property sits, since mortgage broker licensing triggers and exemptions vary by state, and business purpose loans get treated differently from consumer purpose ones in most of them. I'm working through my own licensing now and the difference between those two categories drives half the rulebook. Confirm with the regulator in the state you're lending in rather than taking a broker's word for what applies.

Nobody is pricing their own hours in this thread. A direct shopping exercise on a single bridge file typically eats about eleven hours across two weeks, and that includes the lender who goes silent after collecting the rent roll. If the borrower's time is worth anything at all, ledger's $3,300 gap gets thinner fast.

Cobalt, that is the number worth saying out loud. If shopping costs a borrower ten hours a file, the broker point is really a labor purchase with a side of access. Ledger, on the repeat borrower economics: did any of your brokers cut the fee on the second or third file? A quarter point on a known borrower with clean files still beats zero, and a decent shop should be pricing for retention.

Everyone argues about the point at closing and then the loan lives for nine months. Look at these files from the servicing side and the broker is almost never around after funding. When the draw desk goes silent for two weeks, whoever placed the loan has already been paid.

Trellis, the ones worth keeping do stay in the file. In any service business the follow-through is the only thing anyone remembers. A broker who has closed nine loans with a shop can get a stuck draw released in a day, and a broker who treats funding as the finish line stops returning calls the afternoon the wire hits. Both exist, and the point only earns itself with the first kind.

I thought the broker was paid by the lender and it never came out of the borrower's pocket at all. That's how it was explained to me anyway. Doesn't that make ledger's point column wrong?

Rigging, I had the same idea and I am now less sure of it. Ledger, when the broker was paid by you at closing versus paid by the lender, did the rate move by roughly the amount of the fee? If it did, then it is the same money wearing a different hat and the choice is only about cash at close versus cash over the hold.

The stale list problem is real and it is expensive. Picture a list of four bridge lenders where two change their box inside six months, one dropping anything with a ground up component and one moving from 80 to 70 on purchase advance without telling anybody who is not mid file. The borrower finds out by submitting and getting a no. That is the argument for keeping a broker warm even on deals you could place yourself.

Land is a different planet. Raw acreage with no entitlement, I have never once found a direct lender who'd look at it without an introduction. The broker point isn't a shopping fee there, it's the price of getting the file read at all.

Worth asking on behalf of the readers whose only quotes so far are conventional, fixed rate, with the originator paid by the lender, for whom these numbers might as well be in another currency. Does any of this framework carry over to that side, or is investor debt genuinely a separate market with its own rules on who pays the broker?

Quill, mostly separate, going by what the thread shows. And the whole thread is a fair argument for why so many operators stay out of the flip business. Shopping four lenders per deal is a second job with no title.

I keep two brokers and three direct lines and split by deal type. Cosmetic flip in a market I've done ten in, direct, and I know the answer before I ask. Anything with a use question, a partial vacancy or a seller who needs to close in twelve days, broker, and I don't argue over the point.

Ledger, the thing your sheet can't show is the deals your direct lenders killed that a broker would have saved. Mine's not zero.

Rigging, lender paid compensation exists on the investor side too, and it is not free. The lender prices it into the rate or into an exit fee, so it gets paid across the hold instead of at the table. On business purpose loans the disclosure requirements are much thinner than what appears on an owner occupied consumer loan, and how much has to be disclosed varies by state, so read the fee agreement rather than assuming a standard form exists. The clause worth looking for is whether the broker agreement survives when a borrower closes with a lender that broker introduced and later goes back to that lender directly. Some of those tails run twelve or eighteen months. That is the paragraph that decides whether ledger's direct list is actually his.

The one thing I'd pay a point for every time is a broker who reads a rehab budget correctly. I had a file where the lender's underwriter held back the entire kitchen line because the scope said "refresh" and the budget said $38,000. Scope creep on paper, before a single wall came down. The broker who caught it before submission saved me a month.

Question I'd have been too embarrassed to ask anywhere else. How do you actually know your broker shopped it? I've had two term sheets come back that looked suspiciously like the first one with a different logo. Is there anything you can ask for that proves the file went out more than once?

Compass, ask for the submission list with dates and the declines in writing. A broker who really shopped it has that in their email and hands it over without a fuss. A broker who sent it to one relationship starts explaining why that isn't how it works. Ledger, your comparison is missing exit fees. Term sheets sometimes show a headline rate 0.5 better than the competition while carrying a 1 percent exit fee and a minimum interest period of six months buried two pages later. On a four month flip that minimum interest clause is the entire deal economics. Compare total dollars paid over the realistic hold, not rate plus points.

Worth working out how a broker point shows up on paper. If a borrower pays the broker a point at closing, does it appear as its own line on the settlement statement, or does it get folded into the lender's origination? A question with a boring answer that is still worth knowing before signing.

Down at smaller price points none of this applies. On a $70,000 house a full point is $700 and no broker is getting out of bed for that, and the lenders who'd fund it have minimum loan sizes above the purchase price. Rural small balance is its own gap and brokers rarely solve it.

Hollow, that's genuinely useful to hear because everything I read assumes a $400k loan. Ledger, does the broker math change at all if the loan is $150k? Feels like the fixed work is the same and the point gets much smaller.

I track 22 lenders in a sheet with last contact date, last quoted terms and last known box. About six months in, roughly a third of it is wrong. That's my measured decay rate and it's the honest cost of going direct, because keeping that sheet current is maybe three hours a month of calls that produce nothing.

Which is why I voted the way I did. Ledger, your $3,300 per file gap is real, but you're comparing it against a list you're implicitly maintaining for free.

On the build side, the 0.35 on rate is the least interesting number in this. What matters is whether the lender's inspector shows up in three days or ten, because a crew doesn't wait for a draw and idle labor costs more than the spread. A broker who can say which lender's draw desk actually functions in December is delivering the point that matters.