How do Lenders and GSEs Make Money Selling Mortgages
Lender · Nationwide · Member since 2018 · 571 posts · 310 votes
Hi everyone,
I'm trying to get a better understanding of the mortgage system and all the players in it. To my understanding, a lender issues a mortgage to a homebuyer. They give the buyer a sum of money and receive a mortgage with the house as collateral. Doing this too often would tie up most of the lenders capital, so the lender sells the mortgage to Fannie or Freddie. They then sell the mortgages again on the secondary market in the form of mortgage backed securities.
I'm cloudy on how the lender makes a profit from this model. They don't profit from the interest, since the loan is sold to Fannie or Freddie . Do they sell the loan for a markup, or profit from the administrative and origination fees? And how does Fannie and Freddie make a profit? Why doesn't the lender sell the mortgage as a security themselves and cut out the middle man?
> "Do they sell the loan for a markup, or profit from the administrative and origination fees?"
Yes. They aren't selling a $500k loan for $500k. They might sell it for $515k, and they might flip that same block of money 3 or 7x in a single year.
> "And how does Fannie and Freddie make a profit?"
They are massive. They have trillions of dollars on their books. The P&I payments on trillions of dollars is how they can afford to purchase newly originated loans.
> "Why doesn't the lender sell the mortgage as a security themselves and cut out the middle man?"
Yes, this happens. There's no guarantee that FNMA will be the highest bidder. But when you package these loans up, you will still want it to confirm to the FNMA/FHLMC guidelines, post-2008 we know that the credit rating agencies are trash (every bond is A++++++, just like it was in 2007...), but knowing that it conforms to the FNMA/FHLMC guidelines is a really good baseline. Most of the dog poop loans from 2004-2008 explicitly did NOT conform to either set of guidelines.
When entities are pooling these for sale in a mortgage backed security, what it's comprised of can impact sales price and thus profitability. Maybe the sweet spot is for no more than 10% to be "high balance" loans, or no more than 12.5% in that FICO range of 680 to 720, maybe that is what Wall Street is paying the most for right now. What individual banks will do if they have "too many" of the "wrong" type of loans, is make consumer pricing worse for those ones, and if they are lacking another tranche, great, improve pricing there. We see this all the time, maybe in Q1 some lender has killer pricing for 680-720 FICO, then they get 'too much' of it, and in Q2 they price worse than market for that specific segment. And it cuts both ways, if a MBS is all 760 FICO loans, it's going to have rates that are "too low," so even THAT might be a "problem" when you are going to sell it. Consumers ask me all the time "who will you broker my loan to?" and the answer is "I have no idea, you're going to take 1-6 months to house hunt, the market will move between now and then."
Real example: My Q1 2019 85% to 95% LTV SFR owner occupant go-to bank, as of Q4, prices like complete crap for that segment that they were the kings of in Q1, they very clearly swung too aggressively and got too much of that type of business from me and all the the other mortgage brokers, so we haven't been using them a lot for that particular segment of loans. But now they are calling me and sending me emails to check them out for that same segment in Q1 2020 in a few weeks, so evidently they swung too far in the other direction, and need to get more of that same segment in their pipeline in Q1 2020.
When I was a direct lender, this actually kind of sucked. You think they were giving me a W2 and covering my health insurance so I could broker business elsewhere? Heck no they weren't, so I just had to suck it up and accept that I'd lose business when our particular mortgage bank got "too much" of one loan type and had to price it poorly for the next month or so. Wanting to have my cake and eat it too, rather than sucking anything up, was a big part of why I made the switch.