Monday, February 04, 2008

Kroszner on Proposed Mortgage Regulation

by Tanta on 2/04/2008 05:02:00 PM

Fed Governor Kroszner has delivered a speech on proposed new lending regulations from the Federal Reserve (using the authority granted to them to regulate under HOEPA, which specific authority they've had since the mid-90s, but now that the party's over it's a good time to get tough):

It would apply stricter regulations to higher-priced mortgage loans, which we have defined broadly. We were particularly interested in ensuring that protections remain strong over time as loan products and lending practices change. Our analysis of the data suggested that the troubles in the mortgage market generally arise not from a single practice in isolation, but instead from the complex ways that risk factors and underwriting practices can affect each other, sometimes called "risk layering." Therefore, we have proposed using a loan's annual percentage rate, or APR, to determine whether the loan is covered by stricter regulations.4 Because the APR is closely correlated to risk, the proposed protections would cover loans with higher risks rather than single out particular risk factors or underwriting practices.

With the APR thresholds we have proposed, we expect that the new protections would cover the entire subprime mortgage market and the riskier end of the "near prime" market, the latter also known as the "alt-A" market. Covering part of the alt-A market would anticipate possible actions by lenders to avoid restrictions on subprime loans priced near the threshold. It would also address real risks to consumers in the alt-A segment. This segment grew very rapidly, and it layered risks, such as undocumented income, on top of other risks, such as nontraditional loan structures allowing borrowers to defer paying principal and interest. . . .

The regulations would prohibit a lender from engaging in a pattern or practice of making higher-priced loans based on the value of the borrower's house rather than on the borrower's ability to repay from income, or from assets other than the house. This prohibition is intentionally broad to capture all risks to loan performance and the different ways that these risks can be layered. Moreover, the proposal avoids prescribing quantitative underwriting requirements. For example, the proposal would prohibit a pattern or practice of disregarding the ratio of applicants' income to their debt, but it does not prescribe a maximum ratio because the appropriate number depends heavily on other risk factors, which vary from loan to loan.

At the same time, the proposal does offer specifics. For example, it would create a presumption that a lender had violated the regulations if it engaged in a pattern or practice of failing to underwrite at the fully-indexed rate.5 This presumption is derived from the subprime guidance the agencies issued last year.

It bears emphasis, however, that our proposed regulations would be more robust and comprehensive than the guidance. The regulations would apply to all mortgage lenders, including independent mortgage companies. Guidance, in contrast, does not ensure uniformity of coverage. Moreover, the regulations would be legally enforceable by supervisory and enforcement agencies. Just as important, the regulations, unlike the guidance, would be legally enforceable by consumers. Borrowers who brought timely actions could recover statutory damages for violations, above and beyond any actual damages they suffered.

The proposed requirement to assess repayment ability is intended to protect consumers from abusive practices while maintaining their access to responsible credit. We recognize that satisfying both objectives at the same time is a challenge. The proposed rule's potential for consumer actions, coupled with its careful avoidance of prescribing quantitative underwriting thresholds, could raise compliance and litigation risk. In turn, this could raise the cost of credit for higher-risk borrowers or limit the availability of responsible credit. That is why we have proposed prohibiting a "pattern or practice" of disregarding repayment ability rather than attaching a risk of legal liability to every individual loan that does not perform. This approach is meant to preserve choices for borrowers with shorter or weaker credit records while protecting them from lenders who have a practice of disregarding repayment ability or are found to exhibit a pattern of unaffordable loans. . . .

It is not too early to emphasize that the effectiveness of the final rule will depend critically on effective enforcement. The Federal Reserve will do its part to ensure compliance among the institutions it supervises. We also have been instrumental in launching a pilot project with other federal and state agencies to conduct consumer compliance reviews of non-depository lenders and other industry participants. I am sure we will be aided in these efforts by a new system for registering and tracking mortgage brokers recently launched by the Conference of State Bank Supervisors.
The proposed regulation does extend to all mortgage lenders, not just federally-supervised depositories. However, the general standard in a number of places is whether a lender engages in a "pattern or practice" of problematic lending practices, not any individual instance of it. It is therefore meaningless without a regulatory structure of some sort for each kind of lender that can examine large samples of the lender's loans in a long enough period of time to discern patterns, however those end up being defined. (1 in 100 loans? Every other loan? Something in between? What constitutes a "pattern"?) I am not particularly comforted by the idea that all we have at the moment is a "pilot project" to examine non-depository lenders.

The meatiest prohibitions (limiting stated income and prepayment penalties and requiring tax and insurance escrows) apply only to "higher-priced mortgages." In essence, as long as a lender consistently underprices its risk, it gets to write all the stated-income, prepayment penalty, no-escrow loans it wants to. Theoretically, at least, regulators of federally-insured depositories would be on top of the pricing problem during a safety and soundness examination.

However, that doesn't mean that non-depositories would be subject to examination of risk-based pricing models. Nor does it seem to give those examiners who do look at these things much guidance for when risk is underpriced. The logic, apparently, is that: 1) consumers are not harmed by underpriced risk and 2) consumers are harmed by absolute prohibition on stated income, prepayment penalties, and escrow waivers. It seems to me that 1) can be plausible only if consumers never suffer from things like systemic risk or--just say--RE prices driven up unsustainably by "unqualified buyer" bids, or any other adverse consequence of underpriced mortgage risk. However, number 2) is plausible only if the "harm" to the three or four people who cannot possibly verify repayment ability and yet still ought to get a residential mortgage loan is more likely to bring the economy to its knees than across-the-board prohibition on stated income lending would.

So what would constitute a "higher-priced mortgage"? The proposed threshold for first lien mortgages is an APR 3.00 points over the comparable Treasury security. According to the Fed, that should pick up all subprime and "the riskier end" of the Alt-A market. I remain unconvinced.

"APR" is not equal to the initial (or permanent) contract interest rate on a mortgage loan. It is an attempt to calculate the "true cost" of credit by taking into account points and fees paid up front (including mortgage insurance premiums, where required), as well as future changes in the contract rate (for ARMs). If you had a fixed rate loan with no fees or points charged, the APR would be equal to the contract rate. When there are fees and points, they are amortized over the contractual term of the loan and then added to the interest due. If the loan is an ARM, the APR is calculated by taking into account scheduled adjustments up to the fully-indexed rate (the highest rate possible given the current index value, which is assumed to be unchanged over the life of the loan). For a 30-year loan (of any kind), the comparable Treasury security is the 30-year Treasury bond.

At the moment, the national average rate for a 1-year ARM (no initial fixed period) according to Freddie Mac is 5.05% with 0.70 points. If you assume a standard prime margin of 2.75%, a one-year CMT index (currently 2.71%), a 1.00% origination fee plus another $1,500 in other fees (which is a lot) and a $100,000 loan amount, the APR on the "average" loan is 5.73%. The current yield on the long bond is in the neighborhood of 4.348, so this "average" loan is 1.382 over the comparable security, or well within the tolerance.

I could let you get away with stating income for an additional 100 bps in rate and margin and points (6.05% start rate, 3.75% margin, 1.70% total discount points). That would result in an APR of 6.85%, which is still comfortably within tolerance.

In fact, I could charge you 6.50% on the start rate, 400 bps on the margin, and a whopping 3.00 in discount (plus your 1.00% origination fee and $1,500 in other fees) and still be under the APR threshold (7.26%). I personally can imagine a lot of lender stupidity going on for that much "risk based pricing" (all justified, of course, by a sexy FICO score and a happy appraisal), none of which would get me into regulatory prohibitions. Does this additional yield actually compensate adquately for the amount of risk being taken here? The rating agencies, for one, seem to be telling us that it doesn't.

The trouble here is that the Fed wants some way to quantify "problematic loans," but since it doesn't want to make hard-and-fast rules about prohibited or allowable loan terms, it proxies the matter by pricing. In other words, it takes a certain "historical" correlation between loan risk and APR, declares that APRs of a certain level are "higher-priced," and then prohibits some risk-layering on those loans. This is the model that HOEPA has used since the mid-90s. The idea is supposed to be that competitive forces in the marketplace will prevent any lender from charging the highest fees and points to loans that don't have these risk factors. Loans that do have a lot of risk factors eventually bump up into this price ceiling at which--theoretically--we are no longer just pricing actual risk competitively but overpricing. The result, again in theory, is that lenders won't make certain extremely high-risk loans, because they cannot "price" such a risk without hitting the threshold that requires less risk in the loan. (You could not, for instance, "price" stated income at 3.01% over the 30-year bond. Once the APR on the loan is that high, stated income isn't allowed.)

But nobody is pricing or has ever priced stated income anything near like that in Alt-A. The whole idea of "Alt-A" is that the borrowers get "near prime" rates and points; the ability to do that was always a function of pricing logic that "paid up" for higher FICOs and lower LTVs based on sunny appraisals and sunnier assumptions about home price appreciation. We may not be in that kind of crazy pricing land now, at the moment, but there's nothing here that would stop us from going there again if memories are as short in this regard as they have proven to be in the past.

I might not be quite so cynical about the pricing of stated income loans had I not just read this:
One Oakland woman, who asked not to be identified, explained how she exaggerated her income - with encouragement from her mortgage broker - when she refinanced her home.

"He didn't say anything illegal out loud," she said. "He didn't say 'lie,' he just made a strong suggestion. He said, 'If you made $60,000, we could get you into the lowest interest level of this loan; did you make that much?' I said, 'Um, yes, about that much.' He went clickety clack on his computer and said, 'Are you sure you don't remember any more income, like alimony or consultancies, because if you made $80,000, we could get you into a better loan with a lower interest rate and no prepayment penalty.' It was such a big differential that I felt like I had to lie, I'm lying already so what the heck. I said, 'Come to think of it, you're right, I did have another job that I forgot about.' "
This is your helpful mortgage broker, working tirelessly to get the lowest rate and points for his customers. Of course that's rather a perverse approach to loan pricing--go stated income so your DTI is lower so you can get a better rate--but the proposed regulation won't do anything to fix it that I can see.

Really, this is all more of the weirdness you get when you buy into the "it's a subprime thing" thing. You really do have to believe that the problem is primarily interest-rate related (borrowers are defaulting because they cannot afford their rate resets or payment recasts), not house price-related (borrowers can't afford their houses at just about any rate).