If you search the Internet, congressional testimony, academia or the media for insight into how lenders price residential mortgages you're likely to turn up a mountain of discussion of the yield spread premium paid to brokers by lenders. Maybe a little about the fact that mortgage rates reflect where mortgage securities trade in the bond market, but not much that explains loan pricing for retail customers. Nothing that explains what banks make if they originate the loan internally and don't pay a broker or correspondent to do it for them.

My intention is to fill that hole in common knowledge.

YSP update

Briefly, brokers can be compensated for their services in two ways: in up-front fees (often assessed as points, that is, a percentage of the loan amount) and in a back-end payment from the lender based on both the principal amount of the loan and the interest rate on the loan at closing, commonly referred to as the YSP.

As I said, explanations of how brokers price mortgage loans are easy to find. (My favorite was written by Tanta at calculatedrisk.com, but any number can be found with various obvious key words.) The gist of it is that there is a par rate, below which the borrower would have to pay points (a percentage of the amount borrowed) at closing (buy-down points) and above which cash is rebated to the broker, again expressed as points. The "par" rate is not fixed — it reflects competition between lenders and, one presumes, the yield MBS prices imply investors would pay for a par-priced security.

For example, a rate sheet from a big wholesale lender I found online on Oct. 1 (the same week that Freddie Mac's Primary Mortgage Market Survey spotted 30-year fixed rates at 4.32%) indicated a YSP of 0.025 points for a 4% loan closed within 60 days. Rates are shown in 1/8th point increments, and the next lowest, 3.875%, required an upfront cash payment of 0.6 points (negative YSP as it were). So, the par rate on that day, as on most days, was theoretical and could be estimated by interpolating between these two points. For the same delivery period, the wholesale lender offered to pay brokers 0.525 points for a 4.125% loan, 1.025 points for 4.25%, 1.65 points for 4.5% up to 3.275 points for 5.375%.

For two decades, brokers have been required to disclose any YSP by rules adopted under the Real Estate Settlement Procedures Act of 1974 (RESPA), in both a Good Faith Estimate of closing costs and the HUD-1 settlement statement. The YSP must also be disclosed in the Truth in Lending disclosure. Retail and correspondent lenders make no comparable disclosure.

Brokers can use the YSP to defray origination and closing costs (in fact, I think brokers in the '90s may have first come up with the idea of no-cost loans), but its pretty easy to imagine such a pricing structure would be more likely to give brokers an incentive to steer borrowers into higher cost loans than borrowers' credit standing would indicate. Even before the mortgage bubble, studies by consumer advocates and academics linked YSP-incentivized brokers to predatory and high-cost lending, particularly to credit-impaired, minority and unsophisticated borrowers, often in tandem with prepayment penalties and high-risk loan features such as teaser rates, interest-only periods and negative amortization. The devastating volume of fraudulent mortgages originated through wholesale channels in the last decade (so aptly referred to as the Naughties, the Zips, or the Zilches) is also attributed to the YSP.

Dodd-Frank to the rescue

Broker abuses and YSP excesses have played exceedingly well in the media, on the Web and in congressional hearings, all the more so since the mortgage and housing crash. So we should not be surprised that the Dodd-Frank Act of 2010 appears to have outlawed the practice in SEC. 1403. Prohibition of Steering Incentives (by amending the Truth in Lending Act; historically, regulations under TILA by the Federal Reserve Board have been written and promulgated by the Federal Reserve Board as Regulation Z).

Dodd-Frank directs "the board," in prescribing regulations, to not construe any provision of its amendments to TILA as "(A) permitting any yield spread premium or other similar compensation that would, for any residential mortgage loan, permit the total amount of direct and indirect compensation from all sources permitted to a mortgage originator to vary based on the terms of the loan (other than the amount of the principal); "(B) limiting or affecting the amount of compensation received by a creditor upon the sale of a consummated loan to a subsequent purchaser; …. "

An originator is defined, for purposes of this section as any person who, for direct or indirect compensation, takes a mortgage loan application, assists the consumer in obtaining or applying for the loan or offers or negotiates the loan terms. In other words, a loan officer or broker. This definition should serve to include table-funded loans (funded in the brokers name and simultaneously transferred to the lender, so the borrower thinks the broker is the lender).

Oddly (or predictably) enough, the section directs "the board" to prescribe regulations, apparently having forgotten that it has elsewhere assigned the Consumer Financial Protection Bureau responsibility for major consumer protection laws including TILA. Also, and I hope this is a quibble, though I've nearly blinded myself searching, I cannot find an effective date for issuing such regulations. The closest I come is the effective date for the bureau to assume rule-writing authority. That would be the date Dodd-Frank was enacted.

How brave, how bold

So, Dodd-Frank struck a great blow against an already thoroughly demonized channel of origination. At the same time, it explicitly held the back door open for any lender that can fund from its own pocket to receive a markup when it sells its loans to a wholesaler. In other words, smaller depository institutions and mortgage banks acting as correspondents.

I'm not sure why Dodd-Frank needed to make such a grand display of congressional will. The climate changed for brokers well before Dodd-Frank was enacted. The Federal Reserve Board had a similar rule in the works under TILA a year earlier (the final rule, finessed to reflect Dodd-Frank, was published in the Federal Register Septr 24 and is effective April 1, 2011).

Previous Congresses had been on the case, too. Back in 2008, the Secure and Fair Enforcement Mortgage Licensing Act, part of the Housing and Economic Recovery Act required all nondepository mortgage loan officers to register in a national database, pass national and state exams and participate in 20 hours of approved education before engaging in any mortgage loan origination activities (requirements generally effective this year).

Mortgage players also had turned cold on brokers before Dodd-Frank. Reflecting the atrocious credit performance of broker loans, wholesale originators have steadily been cutting broker channels over the last few years (including two of the very biggest, JPMorgan Chase (JPM: 37.21 -0.75%) and Bank of America (BAC: 7.29 -0.14%); this site tracks closures, mergers, layoffs and channel changes. Likewise, mortgage insurers have been curtailing or cutting off third-party originators since early 2009. Earlier this year, the U.S. Housing and Urban Development Department stopped taking applications from brokers for Federal Housing Administration approval. It will allow them to originate loans only if they are sponsored by an FHA-approved lender, effectively shifting the burden of supervising them to lenders. Likewise, Fannie Mae updated its requirements for lenders managing TPO originations, including a requirement for quarterly rather than annual reviews of loan performance. Freddie Mac published a 65-page guide on wholesale originations best practices, which include checks to its exclusionary list of banned persons and entities.

Under such adverse conditions, the role of brokers in the mortgage market has dwindled. According to Inside Mortgage Finance Publications, they accounted in 2009 accounted for 15% of total loan production, down from a peak of 31% in 2005. However, the Mortgage Bankers Association's Origination Survey data indicates brokers once commanded a much larger share — as much as 50% of all loans and 71% of subprime in the second half of 2005.

Less risk, more profit to the banks

The slack, as brokers leave the field, is of course taken up by retail and correspondent channels, 48% and 37%, respectively, according to IMF Publications. But you must bear in mind that most of these loans are funneled into a too-big-to-fail bank oligarchy. According to IMF Publications, nearly 54% of all 2009 mortgage originations were funded by three financial institutions — Wells Fargo (WFC: 29.60 +1.89%), BofA and JPM Chase.

Think about it. Over half the mortgage loans made last year (and undoubtedly this year) were made by just three humongous commercial banks. Add CitiGroup and Ally Financial to get 61.8%. Back in 2005 — the broker's heyday — five lenders, two of them now defunct, accounted for 47.3%. It took 10 lenders to account for almost 61% of 2005 originations.

It didn't take an act of Congress to squeeze out the mortgage brokers, just good business sense on the part of the oligarchy. No disclosure (so it all goes under the political radar!), more profit, and, for the wholesale lending business, significantly reduced risk.

I couldn't get my hands on a correspondent rate sheet, but I think it's safe to assume the "service release premium" that wholesale lenders pay them for sourcing, underwriting (often "delegated" underwriting) and closing a mortgage loan would be bigger than YSP to a broker for the same loan. (The servicing release premium refers to a sale of both loan and servicing rights.)

But the real joy, the truly enriched pleasure of lending goes to the big retail banks. No surprise who they are. IMF Pubs traces 56% of 2009 retail loan production to the big three — Wells, BofA, JPMorgan Chase (JPM: 37.21 -0.75%). Imagine their pricing power!

The MBS analysts understand it. In their prepayment commentaries, they downgrade the risk of a destructive refinancing wave and speak dispassionately of how sticky mortgage rates are given the decline in other market yields. Here's what I mean by sticky: since early April of this year the theoretical par coupon yield for 30-year Fannie Mae MBS (reflects what investors expect for a par priced new issue MBS) has declined by as much as 140 basis points. (The 10-year Treasury yield fell by about 150 basis points.) By contrast, the Freddie Mac survey sees average 30-year mortgage rates declining just about 100 basis points over the same period.

The analysts explain that lenders keep a rein on mortgage rates to manage the workloads in their pipelines. They are avoiding expensive increases in staff that will require layoffs when the rally ends and loan demand dries up. They are running lean and mean.

Why not? Given oligarchic market shares, they needn't fret about volume. Competitive pressure on rates is minimized. And the more lenders hold mortgage rates up —significantly above the par yield investors demand to own the securities — the more they maximize the profit in a loan. First, they have more premium coupon to sell to investors for more premium over par. Second, because they manage the pipelines through pricing, they don't overwhelm demand and put pressure on mortgage spreads. Third, refinancing is subdued, investors are more comfortable paying a premium for a new MBS and the value of mortgage servicing rights (MSR, an asset) is enhanced.

Don't let me mislead you — the difference between rates on GSE-able mortgages and the yields investors demand also reflects the GSE's adverse market fee and loan level price adjustments (risk-pricing add-ons that can be paid up front or financed by taking a higher interest rate and letting the MBS market fund them). But that is another topic deserving of discussion another time.

Let's take a closer look at the pricing math for retail loan origination. In the week ending Oct. 14, the average 30-year fixed-rate mortgage rate was 4.19%, with an average 0.8 fees/points, according to Freddie Mac's survey. Let's assume that average mortgage is securitized in a Fannie 30-year MBS. Some interest will be withheld from the security to pay the guarantee fee and cover the cost of servicing. We can assume 15 basis point g-fee, and last I looked, the minimum servicing strip was 25 basis points. That would leave 3.79% to pay coupon, but over the decades, the TBA MBS market has standardized to whole and half-coupons. (TBA means pool number to be announced. It is one of the many unique market conventions that have evolved to create the deep, liquid, commoditized MBS market that funds more than 90% of new U.S. home borrowings at present.)

The best execution then would be to sell the 4.19% loan in a 3.5% security. The lender retains 54 basis points of MSR, which are booked as an asset. Market values of servicing are quoted as multiples of the servicing strip. They are also not very public. Based on discussion with insiders, the 54 basis points might be booked in the current favorable (slow refinancing) environment at 1.62 to 2.16 points. A friend at a big trading desk has given me a closing price run for Fannie 30-year from Oct. 12, when lenders were responding to Freddie's survey. The easiest way to hedge the market value of that 4.19% loan is to sell it in a forward security sale, say for November delivery. Fannie 30-year 3.5s for November closed at 101-03 (101.0938). That means the lender nets $1.0938 for every $100 loaned to the borrower, plus the lender collects the average 0.8 points upfront. Depending on where the servicing is booked (and assuming my ball park estimates are reasonable), the lender stands to make 3.5138 to 4.0538 points.

And not a penny of that is disclosed to the borrower in the Good Faith Estimate of closing costs or the settlement statement.

A few last words. Higher mortgage rates go into higher coupon MBS, which trade at larger premiums to par. In fact, given the impediments to refinancing conventional loans in the current environment, the premiums on higher coupon GSE MBS are very generous. The value of MSRs on higher coupon loans is somewhat lower, but I wager that if historical pricing were available for this asset class, they are at all-time highs. Don't tell me there's no incentive to steer retail borrowers into paying higher rates.

NOTE: Linda Lowell writes a regular column, called Kitchen Sink, for HousingWire magazine.

Editor’s note: Linda Lowell is a 20-year-plus veteran of MBS and ABS research at a handful of Wall Street firms. She is currently principal of OffStreet Research LLC.