REQUIRED READING: Albert Einstein is credited with saying that ‘the definition of insanity is doing the same thing over and over and expecting different results.’ The due-diligence process used by the secondary mortgage market is a terrific example of Einstein's definition. For the past 20-plus years, lenders and investment bankers that have sought to securitize loans have made representations and warranties based on a process that has proven, over and over again, to be flawed in epidemic proportions.Â
Despite evidence that the due-diligence process has been ineffective for the past two decades, little has changed. The industry is missing a great opportunity to restructure and reform the process. Doing so would require a new way of thinking about and approaching due diligence, but the rewards would be great in terms of reduced risk, time and cost.
Lawsuits by investors, insurers and other victims of the mortgage meltdown have prompted comprehensive studies on the loans contained in mortgage securitizations. Those studies show that as many as 75% of the loans in the securitizations failed to adhere to the guidelines for which the security was warranted. Yet the due-diligence process that was intended to protect investors continued to provide positive statements about these pools.
Due diligence is defined as the process of reviewing a sample of loan files contained within a pool of loans to be securitized in order to affirm that the loans have been originated in accordance with the guidelines and standards included in the representations and warranties.
What went wrong?
A little bit of history on how the due-diligence process evolved is probably in order. The requirement for conducting these reviews was at the direction of the Securities and Exchange Commission (SEC), which required that a minimum of 10% of the pool be reviewed. The reviews were to be of randomly selected files and reviewed by an independent group, which would then report the findings to the investment bank.
In order to help investment banks meet these requirements, a number of due-diligence firms sprung up. The firms typically employed contract underwriters – who were often not experienced underwriters. Investment banks hired these due-diligence firms and assigned specific file reviews for each of their securitizations.
When the reviews were completed, oftentimes, any loan files from the sample that fell outside the parameters were excluded. Then, other loan files were substituted to ensure the loans in the sample met the underwriting parameters of the pool.
The information on the loan files was then sent to the investment bank. Ideally, if these results showed a significant level of variance from the guidelines, the investment banking firm would then either stop the securitization or, at minimum, change the pricing for the loans.
However, the actual results of these reviews were never made public. Pricing on the pool of loans was set at the time of the bid, and the due-diligence process never had any impact on price – so it was meaningless. Investors and insurers continued to believe that the loan risk was as described and assumed the process was protecting them. Ultimately, it wasn't.
When all was said and done, a lot of people made a lot of money performing reviews that were meaningless in the overall securitization process. The reason the current due-diligence program is flawed may be traced to four key problems within the process itself.
First, the sample of loans selected for review is not random. The review process calls for a random sample of loans. By definition, random sampling should be the selection of items – in this case, loan files – from a population that ensures each item has an equal chance of being selected. The process is intended to evaluate the probability that the errors found in the sample population are found in the entire population, and to do so within a specified confidence interval.
However, if loans are selected for review based on specific criteria not shared with the entire population, the sample is biased, and the results are not applicable to the entire population. The selection processes for these files typically included some random selections but were largely made up of what the investment bank perceived to be higher risk.Â
For example, the investment bank may have decided that Arizona loans were riskier, so a large percentage of the sample was made up of Arizona properties – even though Arizona may have made up less than 2% of the entire group of loans included in the securitization. If they found this assumption was wrong, they made a statement that the pool was acceptable, even though there were many loans that did not meet the guidelines.Â
Second, there is no standard methodology for reviewing the selected loans. As a result, there is no way to compare populations or samples, and no way to differentiate between high-quality and low-quality product.
These reviews were intended to provide investment banking firms and others in the business of securitizing pools with the assurance that the loans were quality loans. However, this approach was doomed from the beginning, because there is no standard definition of what constitutes a quality loan. All too often, reviewers seemed to rely on the old adage, ‘We may not be able to define it, but we'll know it when we see it.’ Unfortunately, everyone had a different vision.
Third, the due-diligence process is highly subjective and varies between due-diligence firms. Each due-diligence entity has its own approach to the review process, with its own focus for reviewing risk. As a result, the reviews are biased.
As demonstrated many times since 2008, this subjectivity was not always the result of the individual reviewer's analysis, but was often based on what was expected by the firm's management. In some cases, it was the result of an investment banker's influence.  Â
Finally, due-diligence reviews were often contracted out to personnel who had little or no underwriting experience. The individuals conducting these reviews were primarily contract personnel. In order to minimize overhead costs, the due-diligence firms maintained a minimal number of full-time employees and kept a database of individuals who could be called upon to participate in the loan file reviews. While some of these individuals were highly experienced and knowledgeable, many actually had limited or no experience in underwriting.
Fixing the errors
At the end of the day, the process for ensuring the quality of the loans was driven by the due-diligence companies' idea of what was meant by quality, not the expectation of quality the investors and insurers wanted.Â
Is there any doubt in anyone's mind that the Quality Residential Mortgage and Quality Mortgage requirements now being hotly contested on Capitol Hill are the result of anything other than the failure of our current due-diligence process? Why else would lenders be required to have some ‘skin in the game’ except to try and force them to be diligent about how they produce these loans?Â
In order for due diligence to be a meaningful process, some radical changes are needed. The quality of the product must be equated with the level of operational risk inherent in each lender's processes. Those processes must be clearly identified and priced in each loan, and in each security.
For that to happen, due-diligence reviews must incorporate four major changes. First, there must be industry and regulatory acknowledgment of the impact of operational risk. We must be willing to identify and price for the risk that the guidelines were not followed, that mistakes were made and that this will ultimately cause a greater likelihood of default. Blaming the mortgage meltdown on product and documentation is far too simplistic for these very complex transactions.
Next, a statement of the overall rate of unacceptable loans in the sample and an inferential statement as to the confidence level of these results for the entire pool of loans must be provided with each pool analysis. This means that each loan must have a yes or no answer to the question, ‘Does the quality of this loan meet the expectation of the investor?’ It does not mean there is just a list of findings that fail to establish the overall operational risk associated with the pool.
Then, the reviews must be consistent and conducted according to the quantifiable risk associated with individual process variances, as well as any associated layering of these risks. The initial quantification of these operational risks has been completed for some time, yet firms continue to use the same approach to reviewing loans and identifying problems, while rejecting the more reliable method that has proven useful in numerous other industries.
Finally, this risk analysis must be incorporated as an independent, automated part of each lender's quality-control program. Why should the industry continue to pay millions of dollars for a program that does not work when a valid, meaningful analysis can be conducted during the origination process for a minimal cost?
This would allow the lenders to fix mistakes prior to funding, as well as allow the secondary market to price for the risk. And, they could do it for pennies on the dollar. Recent changes by Fannie Mae and Freddie Mac incorporate a small piece of this type of analysis but lack the depth of review and fail to accept the quantification of the risks that are an inescapable result of the product's performance. Â
There is no question that some means of ensuring the soundness of loans produced by the industry will ultimately be included within the myriad of new regulations with which we are faced. While the focus today is simply trying to come to grips with integrating these requirements and maintaining some level of profitability, we seem to be missing the opportunity this crisis has provided. Â
It's time for the industry to stop the insanity, be smart and make residential mortgage-backed securities a trusted and valuable investment vehicle.Â
Becky Walzak is president of the Indianapolis-based Looking Glass Group LLC. She can be reached at (561) 459-7070.
