Know before you go

Brian Levy is the guy I call when I’m trying to disrupt something and want to make sure the shit doesn’t hit the fan.

He’s Of Counsel at Katten & Temple, LLC, and he’s who a lot of mortgage CEOs call when they need a straight answer and want to stay out of trouble. Before private practice, he spent 15 years as general counsel of a bank that owned three mortgage companies (retail, wholesale, and home equity lending), managing loan officers and origination joint ventures. He’s seen the business from inside the Board room, sales floor, and from inside the courtroom. He also writes Levy’s Mortgage Musings, one of my favorite newsletters in the industry, because he says what he actually thinks about regulators and the business when most people in his seat keep it vague. I subscribe to and read it every month when I get his email from Substack.

So I interviewed Brian for a reason. He spends his days advising companies. Companies like mine. Possibly the company whose offer letter is sitting in your inbox right now. Originators almost never get access to someone in his seat, and I wanted to fix that. I wanted him talking to the originator for once, not the company.

Because the more you know, the more unleashed you become. Not because somebody handed you the answer. Because you finally have enough of the picture to think for yourself.

One thing first, and Brian asked me to say it, so I’ll say it plainly. Nothing here is legal advice, from him or from me. It’s a conversation between two people who’ve been in this a long time. Read it that way.

Here’s where it starts. Every originator who has ever changed companies has done the same thing. You get the offer. You go straight to the comp plan. Basis points, signing bonus, the split. You read that part twice. Then you skim the rest, sign it, and start.

The part you skimmed is the part that decides how it ends. What happens to your pipeline when you leave. Who owns the database. Whether your pre-approvals can move with you. You don’t find out the agreement had a whole second purpose until the day you try to walk out, and by then it isn’t a negotiation. It’s just the rules.

When I asked Brian about this, he told me something that stuck. In all those years, he can’t remember a single originator ever coming to him to read an agreement before they signed it. The companies pay him. The companies have always paid him. The originator signs whatever shows up, usually without an attorney looking at it.

Then he said four words I haven’t been able to shake.

“And maybe I should.”

The agreement is built to protect them, not you

Start with the document itself, because when Brian drafts an employment agreement for a company, he’s given exactly one job. “It is rare when someone talks to me about that,” he told me, about whether the agreement should be fair or reflect a culture of cooperation. “They want to make sure that if we end up in court, we’re going to win.” That’s not a scandal. It’s the assignment, and often for many companies, Levy says, an assignment born out of having been burned in the past by mercenary loan officers. So, a company hires an expert to make the document protect the company, and Brian is good at his job. What’s unusual is a company-side lawyer willing to say it out loud to the people on the other side of the table.

So read it knowing what it was built to do. And read it going in, not going out. That’s the whole title of this thing, and Brian said it more than once.

“The time to know that is when you’re going in, not when you’re going out.”

Almost everybody does the opposite. They read the comp number on the way in and the fine print on the way out, by which point the terms aren’t terms anymore. They’re just facts you’re stuck with. The leverage was all at the front, and most originators spend it without ever knowing they had it.

I couldn’t get Levy to get specific about the exit details, but I have no problem discussing the obvious questions to ask. Who owns the customer database when you go. What happens to the loans in your pipeline, and at what point they become the company’s. Whether your pre-approvals can move with you or are trapped. Whether the non-solicitation covers borrowers, employees, or both, and whether it binds people you recruit if you ever build a team. Levy reminded me too that it’s critical to know what you’re carrying in from your last shop, because you have two agreements to worry about, not one. The old one still binds you, and the time to read it was before you started interviewing.

Then there’s the part where the agreement stops mattering and your behavior takes over. When you leave, the rule is simpler than people make it. Don’t be sneaky and treat your old company fairly. Don’t hold back files. Don’t take borrower personal financial information out of the company’s systems. Your contact list, the one already in your phone, is probably yours, but is it your phone? The company’s CRM is not. And don’t think you’re clever emailing it to your home account. As Brian put it, “it is easy today to track what you’re doing within those systems,” down to the computer you’re typing on. The cleanest agreement in the world won’t help you if what you actually did on the way out the door is sitting in a log file. Frankly, Levy says, a lot of litigation in this area could be avoided if folks just were honest with each other and worked together to make sure the transition was handled professionally in the best interest of the consumers.

My advice: get the recruiting promises in writing. If a recruiter promises you a corporate margin, 50 bps and you keep the rest, the promise is worth exactly nothing until it’s on paper. Whatever they committed to that affects your pay, you want it in the contract. Then ask for a way to verify it. Not a full forensic audit. Just a mechanism to confirm the margins, costs and allocations are what they said.

Because here’s what he told me, unprompted, about why those promises never make it onto paper. “Probably because people like me that represent those companies say, no, you don’t want to put that in writing, because you don’t want to have any LO get contractual rights over your ability to make changes to your program.” What that means is that six months in, when you call for a price check and the margin has quietly drifted, there’s nothing that says they did anything wrong.

A company comfortable with what it promised should be fine showing you how to verify it. A company that suddenly gets vague is telling you something too. The contract said one thing. What they actually do with your margin is the thing.

Once he said it about agreements, I heard it everywhere.

Once Brian said it about employment agreements, I started hearing it everywhere else in the business. The same mistake, over and over. People think the paper saves them. It doesn’t. The paper is a structure. What you actually do inside that structure is what holds up or doesn’t.

He put it plainly, talking about joint ventures as a way to compensate referral sources. “We could just create a joint venture people say. No, you actually have to have a joint venture. You can’t just put something on paper and say, well, it was a joint venture, so that’s why we were paying the referral source.” The paper (hopefully) just describes a legal structure. Even if what you have on paper is legit, it does not save you from what you actually did.

“What you do is what matters.” He said that, almost word for word, about half a dozen different things in an hour. Hold onto it. Everything below is the same idea wearing different clothes and Levy’s Mortgage Musings blog has been talking about these things for years.

The schemes that are “legal” until they aren’t

You’ve seen the pitches on LinkedIn. Get a realtor licensed as an originator and pay them on referrals. Hire the real estate broker owner as a non-producing sales manager and pay them on profits. Skip the MSA, skip the JV, just put them on the payroll. Easy money, the attorney signed off.

Brian’s been watching this exact idea since the early 2000s. He pointed to a HUD settlement back then involving a company called Z-Net, where the government’s whole argument was that the employment was on paper only, that no real employment was actually happening and realtors were getting paid for names on a napkin. Twenty-something years later, it’s the same idea with a new coat of paint, and it usually falls down in the same place.

For any of it to work, Levy says, the person has to be a genuine W-2 employee. There’s no minimum number of services they have to perform. You can be the name on the billboard who hasn’t taken an application in five years and still get paid, as long as you’re truly an employee. But that’s the whole ballgame. “If they’re not really your employee,” Levy said, “then what are you paying them for?”

So what makes them a real employee? In true lawyer fashion, Levy says “it’s complicated”, and “there’s no clear test”, but it’s definitely not just a clause in a contract. It’s the actual job. They have to be trained (including compliance). They have to be supervised. They have to be licensed if they ever touch a consumer. They have to go to the meetings, report who they called on and what they did, sit under a manager, get paid on a W-2, and do the same things every other originator at the company has to do to get paid. That is the “rubber meets the road” test. Not the agreement. The behavior the agreement is supposed to describe.

Here’s where that idea dies. Most realtors and broker owners don’t really want to do any of that work. They’re 1099 by temperament. The meetings, the supervisor, the reporting, the licensing, none of it is in their DNA, and that’s exactly the point. The moment they won’t do the work, the RESPA “employee/employer” exception the whole structure depends on evaporates. Now they’re just getting paid for a referral. Now it’s a RESPA violation. Now you’ve got class actions, regulators, treble damages, and legal fees that run into the millions even on the cases that settle.

So I pushed him. Can it be done legally? Yes. Many of the arrangements that have the paperwork, are they actually compliant in practice? That’s a different question. I asked where it unravels.

“It’s actually not as challenging from a legal standpoint as it is a practical and, you know, just execution (and evidentiary) standpoint.”

There it was. The whole thing in one line. He kept going.

“Folks think if you have something on paper a lawyer wrote that they’ve cracked the code on RESPA and can just start paying their referral sources. But, executing is the hard part. If you can’t execute, then don’t bother. And frankly, what I’m telling you is whatever the idea is, most folks can’t execute it.”

The 1099 broker model spreading out of the West Coast has the same disease plus a second one. Everybody checked whether their state allows 1099 status, then stopped. Nobody checked RESPA. Think it through. If you’re a 1099, you’re an independent contractor, which means you don’t get the employment exception to RESPA at all. So either you’re a broker performing your share of the services, or you’re getting paid for a referral. That’s the first problem.

The second one is worse. If the broker shop gets paid by the lender to broker the loan, and then splits that fee with a 1099 originator, that’s a split of a settlement service fee. That’s Section 8(b) of RESPA, the illegal split. “As I noted in a 2023 Musing called “1099 Employee or a duck?”, I don’t see how the 1099 broker gets around 8(b),” Brian said. “The licensing doesn’t help them at all from the RESPA perspective. It just gives them a license.”

So, let’s be brutally clear about it, because this is the part people skate past. The structure is the easy part. Anyone can paper a structure. The supervision, the licensing, the training, the documentation, the meetings, the reporting, the actual bona fide work, that is the hard part, and that is the part almost nobody does. A perfectly drafted agreement that an attorney blessed does not protect you if what happens day to day doesn’t match the page. “Our attorney signed off” is not the same sentence as “we actually run this as real employment.” Until the second sentence is true, the first one is just decoration.

The one place where the rule itself is the problem

I want to flag one exception. On the schemes above, Brian’s answer is always execute better or don’t bother. There’s one area where he doesn’t say that. He says the rule itself is bad policy.

LO Comp.

I asked him how he’d design the LO Compensation Rule. “I’d get rid of it entirely.” Then he told me why.

“It’s anti-consumer, the idea that you have to fix a price, that at the point of sale somebody can’t lower their own compensation. It’s just anti-consumer. It’s ridiculous and the penalties for violation apply regardless of consumer harm.” Again, Brian pointed to multiple blog posts he’s written on the subject, most recently last year in “The LO Comp Rule is Brutal”.

The LO Comp rule was built to stop a specific abuse. Originators steering borrowers into subprime garbage loans to fatten their own check. That was a real problem in 2006. It is not the problem anymore. The subprime products are gone. Between the QM cap, ability-to-repay, and fair lending, the market that the rule was designed to police has already been policed by three other things. What’s left is a rule that forbids an originator from cutting their own commission to save a borrower money on a deal. That’s the part Brian calls ridiculous, and he’s right. A rule designed to protect consumers now stops an originator from doing the most pro-consumer thing available to them.

But here’s where Brian stays honest instead of just venting, and it’s why I trust him. I asked what happens if it’s abolished. He didn’t sell me the fantasy.

“That’s not going to happen. Because not only do you need to get rid of the rule, you need to get rid of sections of TILA, which were passed in Dodd-Frank.” Congress would have to amend the statute. This Congress isn’t going to do that. So full repeal is a daydream.

What he actually expects is narrower and more real. Targeted relief. The ability for an originator to reduce their own compensation at the point of sale. Some flexibility on bond loans and CRA loans, where there’s so little revenue that a small commission cut could be the difference between doing the deal and walking. And fixing the chargeback rule, which right now is almost exactly backwards. You can charge back an error nobody could have foreseen, but not one that was foreseeable, which is upside down, because the foreseeable mistake is the one somebody actually blew.

That’s the difference between a guy with a grievance and a guy worth reading. He’ll tell you the rule is stupid. Then he’ll tell you it’s not going anywhere, and exactly which pieces might actually move.

And no, you can’t outsource it to the robot either

Same lesson, one more time, with the newest tool. Originators are building their own AI tools, running borrower conversations through them, ingesting Social Security numbers and income docs to spin up pre-approvals. Brian’s view of AI is the calmest one I’ve heard from anybody in his seat, and it’s the same idea again.

“You can’t blame AI. It’s just a tool.” Like a calculator. If it comes back wrong, that’s on you, not the machine. He uses it himself, and says he has to keep telling Claude to stop agreeing with him, because what he wants is the holes, not the applause. Supervise it like you’d supervise a sharp subordinate who’s occasionally confidently wrong, and you’ll get a lot out of it.

He had one hard line. Don’t run borrower data through public AI tools. That is borrower financial information your company is obligated to protect under Gramm-Leach-Bliley, state privacy law, and internal policy. Nobody fully knows yet what that liability looks like when it leaks. Someone is going to be the test case. Don’t be that person. The tool didn’t violate anybody’s privacy. You did.

What he’d never been asked to design

Near the end, I described how I write my own agreements. I want them middle of the road, the version two attorneys would land on if both were in the room. Clear non-solicitation. Clear pipeline terms, so that if someone leaves we’re not fighting over loans. We pay departing originators on loans that close after they go, specifically so nobody’s incentivized to grab a file out of underwriting on the way out the door. Punish people for leaving and they behave badly leaving.

Brian’s response stayed with me. “That is a very unusual assignment for an attorney.” Then, “I have never seen it done your way.” He said he’d love to put terms like that in more agreements. And then he told me why he can’t.

“Most employers want to put a written handcuff on somebody that makes it difficult to exit in the hope that either they will think twice about leaving or, at a minimum, will do so in an orderly and professional way without the sneaky stuff on the way out.” As with any relationship, over time, trust can grow or wane, but the agreement doesn’t change. If the relationship is good, however, even on the way out it shouldn’t matter what’s in the agreement.

Which is the whole thing, isn’t it. The handcuff is the paper. The relationship is what actually happens. You can have a beautiful agreement and a company that treats people like prisoners, or a fair deal that two people actually honor.

Go read and subscribe to Brian Levy’s Mortgage Musings. I get it delivered to my email every month.

And the next time a recruiter slides an offer across the table and calls it standard, you know the only question that matters now. Does the contract codify the promises made during the recruitment process?

And be careful of “the lawyer signed off” pitches. Regulators don’t care what the paperwork says. They care about what actually happened.

Rich Weidel
CEO, Princeton Mortgage