Sunday, April 4, 2010


Dear friends.

With changing life circumstances, it will become more difficult to post with any regularity. I welcome all comments, but please accept my apologies in advance for not being able to respond in a timely manner.

Best of the year to all.

Wednesday, March 24, 2010

A Not-so Gentle Nudge to Restart Mortgage Securitization

The financial regulatory reform bill is speeding through congress.
There are many sections which represent progress, but I fear the portion dealing with Securitization would not do enough to reassure future investors. I shall review what's been proposed and discuss their shortcomings and propose an alternative which may help restart this market segment.

I. Excerpt from the summary of the revised financial regulatory reform bill unveiled Monday

(3/15/10) by Senate Banking Chairman Chris Dodd:

Companies that sell products like mortgage-backed securities are required to retain a portion of the risk to ensure they won't sell garbage to investors, because they have to keep some of it for themselves.

Why Change Is Needed: Companies made risky investments, such as selling mortgages to people they knew could not afford to pay them, and then packaged those investments together, called asset-backed securities, and sold them to investors who didn't understand the risk they were taking. For the company that made, packaged and sold the loan, it wasn't important if the loans were never repaid as long as they were able to sell the loan at a profit before problems started. This led to the subprime mortgage mess that helped to bring down the economy.

Reducing Risks Posed by Securities

- Skin in the Game: Requires companies that sell products like mortgage-backed securities to retain at least 5% of the credit risk, unless the underlying loans meet standards that reduce riskiness. That way if the investment doesn't pan out, the company that packaged and sold the investment would lose out right along with the people they sold it to.

- Better Disclosure: Requires issuers to disclose more information about the underlying assets and to analyze the quality of the underlying assets."

Mortgage Bankers Association (MBA) chairman Robert Story Jr. suggested that qualified residential loans with certain characteristics like a 30-year fixed rate, full documentation and a sufficient downpayment should be exempt from the risk retention requirement.

This idea is actually reasonable as the conforming loans have been performed well and the oversights from Fannie Mae and Freddie Mac do maintain a minimal underwriting standard. However, I fear that we're focusing on the wrong questions in the current legislation.

There are several myths which needs to be dispelled to get a clear picture toward possible solutions. Having worked at major issuers of conforming and private label MBSs and ABSs, I have detailed knowledge of the securitization process and I shall endeavor to bust some of these myths.

1. Risk Rentension

While the risk rentention proposal is a sensible clause, it wouldn't have prevented the mortgage credit crisis. This is becuase many of the alt-a and subprime issuers ALREADY retained a portion of the mortgage risk. The buyers have been knowledegable enough to demand the issuers to 'retain skin in the game' for many years already.

2. Capital backing the residual risks

Again this was already done - many issuers did set aside reserve to cover potential losses. The problem was the housing price assumptions being used.

3. If you pool together junk, you can't make an AAA asset out of it.

While this is true in a broad sense, credit support structuring in MBSs and ABSs do work to a degree.

Consider this analogy. Say you have one hundred 90lb weaklings who can each lift 50lb each. While none of they can lift 2500lb individuall, 50 of them together CAN accomplish this feat. Another 25 of them can lift 1250lb, etc. The general idea of MBS/ABS is to pool cashflow from many mortgages to support multiple classes of securities. The most senior class get the support of everybody in the pool. THe second most senior class get the support of a smaller portion, say 60% of the pool, etc. It is in this sense that the rating agc give the most senior class a AAA rating: that it could withstand a high % of defaults among the mortgages within the collateral pool and still be able to repay its principal.

So the future value of this senior class depends on how many collateral mortgages default. If the default stay below a designed %, then its value is not impaired. The defaults hurt the less senior classes. If defaults trend above this level, then the value of the senior class is diminished.

From the above discussion, you can see that the assumptions used for the expected default and loss are critical parameters. With consistent underwriting on a tested product, one can arrive at reasonable assumptions - i.e. conforming loans are performing reasonably well.

The key surprise was the severe home price drop which was beyond the range of most model assumptions. The system failed to adjust and reprice the securities according to the increasing amount of risk that was becoming apparent in 2007-2008.

This failure is due to severe misalignment in incentive structures which CAN be fixed via regulations or best practices. Having observed these misalignments in industry, I have found a good vocabulary to talk about them in the book "Nudge: Improving Decisions About Health, Wealth, and Happiness" by Richard H. Thaler and Cass R. Sunstein.

II. Structural misalignments

As a general principal, when the full costs and benefits of a decision is NOT fed back to the decision maker, but are separately directed to distinct parties, this leads to mispricing within the decision.

The current securitisation process suffers from 2 separations and 1 darwinian effect which cause mispricing.

1. Separation in Time

Pool mortgages; Structure; Sell rated classes now; Get bonus.
If the structure loses money later, it's not my problem.

2. Separation in Space (organizational groups)

Performance bonuses go to securitization/structuring/trading departments and the Executives.

The retained residual (equity class) and low rated classes are usually managed by the asset management department. If the residuals and low rated classes lose money later, mostly the asset management department lose out on their bonuses. If there are enough securitization deals occuring, the income overwhelms the loss from the asset management area, the Executive bonuses won't be affected much.

3. Red Queen or Chuck Prince effect

"When the music stops, in terms of liquidity, things will be complicated. But as long as the music is playing, you've got to get up and dance. We're still dancing," CitiGroup's Chuck Prince said in an interview with the FT in Japan.

Once having set up the securitzation pipeline, staff, IT, etc. a firm wants to keep them working, so it would pursue deal which may not be profitalbe in the long run. As more dealer pursue the same securitization pie, the deals get worse as the collateral underwriting get looser. This is related to Separation 1 above as the increased risk is not apparent now.

Without a mechanism for resolving these structural problems within the securitization process, the reform bill would not prevent a future recurrences of credit crises. Others have suggested deferred compensation and clawback provisions, but that can be complex to implement and would not provide much feedback.

III. A Simple Proposal

"Require that a portion of the bonus for people sponsoring securitization be paid in form of the least senior classes (BB and below) and the residuals created within that year." These sponsors include Executives, structurer, banker, trader, underwriter, etc.

Such classes are inherent illiquid and rarely sold. They usually only pay down as mortgage principal is paid off.

In this scheme, if the residual is overpriced to make a deal go through, the structurer would be hurt by lower cashflow in the future. If they underprice the residual, then the income is lower for the current year and they would get a smaller bonus. If the underwriting is too loose, one may get higher current income, but would lose out on future bonus cash flows.

This naturally ties together the costs and benefits of the choice for doing the securitization and for making the loans. Perhaps even more important, it would make the quality of the loans/classes 'Salient' to the sponsors. While a bonus is usually granted once a year, the bonus cash flow would be reported monthly. Whether the mortgage collaterals are performing well or badly, the impact would be salient (in their face) to the sponsors each and every month.

The loan origination and securitization firms can benefit by not having to set aside reserve against classes that they retained, if they are put inside this bonus pool since the risk is now transferred from the firm to the risk creator. If there are not enough residuals and lower rated classes, the firm can fund a separate bonus fund with payout indexed to the residual pool paydowns.

Some of my colleagues in the securitization field may resent this proposal. I would only remind them that until the investor regain her/his confidence in the nderwriting and securitization process, NO ONE will be making any money from securitization. And that would be a shame since I consider it a most worthwhile tool when used in the right way.
All contents on this blog copyrighted by respective authors (2008). No rights have been conferred to any other parties by the act of publication. You are most welcome to link to this blog or make fair use of it.