Between 1989 and today, securitization markets, and therefore the capital markets, have replaced banks as the lead funding for home mortgages. It is true that excessive social engineering to over-stimulate housing purchase drove speculation. But in my view, poorly developed and opaque securitization market, drove excessive liquidity and irresponsible lending and borrowing. Without the confluence of these issues we would not have had the withdrawal of liquidity to the mortgage finance market and an ongoing cycle of falling home prices. This opacity is the actual root of the crisis, and it led to the ultimate breakdown of the private securitization market.
Today, as it was in the prelude to the crisis, securitization markets too often operate in a “Wild West” environment where the rules are more often opaque than clear, standards vary, and useful and timely disclosures of the performance of loan level collateral is hard to come by. Asymmetry of information, between buyer and seller, is the standard.
Current problems in the real economy, stemming from the opacity and information asymmetry of the asset backed securities (ABS) market, are not isolated to private first-lien residential mortgage securitization markets. They extend to other areas of consumer financing, like home equity, cards, and auto. They also involve commercial financing, like commercial mortgages, construction loans, bank trust preferred, corporate loans, and commercial paper. However, because of the excessive degradation of mortgage underwriting standards and the growth in mortgage funding, we have seen the earliest and most serious damage in this sector. Consider the scale of this growth: between 1985 and 2007 the ABS market grew dramatically, from $1 billion in new issues to $997 billion in new issues. (See Figure 1.)
To believe that real estate or the economy itself can find a self-sustaining recovery without first repairing this important tool of financial intermediation is unrealistic. Liquidity cannot efficiently find its intended target unless there are credible markets in which participants can foster financial intermediation and through which capital can be transmitted. Expanding the monetary base without an effective means of financial intermediation can result in little more than hoarding. Other than fostering new asset bubbles, it may have little sustainable productive economic impact.
Repair and Restructuring
Since 2007, those parts of the securitization market that are not fully subject to implicit and explicit subsidies or guarantees by governments, or do not have robust standards, have ground to a virtual halt. We must set about to fundamentally repair them. These repairs are achievable, but they must be real and fundamental. They cannot be merely another iteration of the same flawed market with the same skewed incentives. Investors –- the key intermediaries in capital formation — need to lead the redesign. They cannot be subject to information asymmetries, fee-arbitrage opportunities and other structural flaws imbedded in the issuer-led design of the prior securitization markets.
While the economy has, for the moment, exited recession, the risk remains that without functioning securitization markets, many of the credit constrained assets formerly funded by securitization will continue to follow the housing market in collateral value declines.1
If it is correct that the real economy problems with housing are not the root of the crisis, then many of the problems in the real economy which stem from contraction in credit availability may be symptomatic of securitization market failures. There is an immediate need for regulators and policymakers to oversee the creation of a standardized market where assets can be securitized, priced, valued and consistently evaluated by investors. In recreating the structured market, we must also clear outstanding legal questions2,3,4 about matters such as “true sale”.5,6 Without clarifying the clear legal and accounting standards on “true sale”, issuers of a securitization may retain rights to or responsibility for collateral that they thought they sold and the investor in a pool believed himself to have purchased.
The primary market for securitizations is different from the equity markets. There is no “red herring” or pre-issuance road-show period during which investors have the ability to really analyze a deal and its underlying collateral. Typically, deals come to market so quickly that investors are forced to rely on rating agency pre-issuance circulars, term-sheets or weighted average collateral data. These tools have proven inadequate.
In order to accurately price securities, investors need timely loan-level performance data on the assets backing each deal. We need loan-level data on a daily, or at least monthly, basis in both the primary and secondary markets.
Without frequently updated and standardized disclosure of loan-level data, market participants can’t independently analyze and credibly value asset-backed securities based on full information. Previously, investors didn’t know what they were buying. Currently, investors are staying away from the securitization market. A massive withdrawal of funding to key parts of our economy is the unfortunate result.
The parts of the private issuer securitization market that are governed by the SEC’s “Regulation AB” are currently functioning more fully than those not subject to reporting and disclosure requirements of “Reg AB”. Where many non-revolving collateral class securitizations have ground to a complete halt, credit card and auto securitizations continue to function, though some at a lower level because of concerns about the quality of consumer credit in the real economy.
The Need for Disclosure
To ensure adequate transparency, enhanced disclosure rules should be required both for deals with and without static pool data (such as asset backed commercial paper). Data on the specific underlying collateral in each pool should be made available for a reasonable period (not less than two-weeks) before a deal is sold and brought to market. This should be done to enhance investor due diligence, to foster the development of independent analytical data providers, and to reduce reliance on rating agencies, The loan-level data should be available in an electronically manageable and standardized format.
While full elimination of the rating agencies may or may not be necessary or realistic, in my opinion we must reduce reliance on ratings and support a narrowing spread between price and value in the secondary market. To that end, the SEC should require that after the deal is sold, all data fields in the pre-issuance disclosures and material information about the loan level collateral in the pool should be updated and be similarly disclosed on a daily, or at least monthly, basis in an electronically manageable and standardized format. Regardless of the nature of the deal (private placement or registered) the data should be publicly disclosed to the loan level and all servicer advances to the pool shall be disclosed as such on a timely basis. Any subsequent repayments of servicer advances should also be reported in a clear manner.
Capital and markets would be less volatile if they could fully model the expected performance of underlying loan level collateral data before a deal comes to market and, on a regular basis, reassess the deviance from expectation. The regular and timely updates to collateral data would reduce volatility, since degradation in a pool would be observable and thus priced in over incremental periods. By requiring that investors receive early and regular disclosures of all available data and adequate levels of information about the underlying collateral, the importance of rating agencies’ recommendations will be diminished to the level of an equity analysts’ research note.
Rather than recognize this lack of timely loan-level performance disclosure standards, regulators and legislators have been pushing to require issuers to hold a slice of every deal they issue. On the surface, this appears to make sense. But on closer examination, that requirement would not have prevented the past crisis and it probably won’t prevent the next one. Many of the firms that have been harmed by holding these securities were the same firms that issued them. The retention argument comes from the belief that issuers may have knowingly sold toxic securities. But more often, these firms didn’t have the available information or resulting ability to fully model their exposures. To force them to increase concentrations of these held securities will only increase their risks.
If detailed loan-level performance data were provided, investors could properly analyze risks to the pool. In that environment, prohibiting retention would actually reduce the risks to our regulated financial institutions, because the problems faced by Merrill, Bear, and others resulted directly from retained exposures to tranches of securitizations they thought were appropriately risk modeled — and turned out not to have been.
In the lead-up to the crisis, even primary financial regulators could not analyze or even have access to deal documents of CDOs their regulated institutions held.7 The automation, standardization, and public disclosure of key collateral information before a securitiztion is marketed — and at least monthly after it is sold — is a necessary ingredient to the development of the deep and broad markets necessary to fund our economy.
In further support the ongoing development of deep and broad markets and reduce the gaming of mark-to-market values, the SEC should require that, on a daily basis, all dealers publically disclose the last trade prices of all ABS, regardless of whether they are otc, bespoke or registered.
Contracts that Work
We also need to address the lack of uniformity in the contractual obligations between various parties to a securitization. “Pooling and Servicing Agreements” (PSAs) and “Representations and Warranty” terms can be several hundred pages long. They define features like the rights to put back loans that had underwriting flaws, the responsibilities of servicers, and the relationship between the different tranches. In addition, key terms that define contractual obligations are not standardized across the industry, across issuers of securities with the same type of collateral (e.g. RMBS, CMBS or RMBS based CDOs) or even by issuer (each issuer often had several different Pooling and Servicing Agreements and Representation and Warranty Agreements).
The lack of standardization and the length of the documentation effectively created opacity, which contributed to the problems in the securitization market. When panic set in and investors began to question the value of their securities, they knew that they did not have the time to read all of the different several-hundred page deal agreements. This reinforced the rush to liquidate positions. What investor wants to be the last one holding a security whose terms he doesn’t fully understand?
This “run on the market” caused securities’ values to fall further than fundamentals would have justified. But without clarity of contract or sufficient, frequently updated, loan-level information to readily analyze the underlying collateral values, there was no other possible outcome. As a result, even investors that focused on distressed securities could not identify, analyze and invest in these securities in the timely manner necessary to provide a floor under prices.
The industry has only recently moved to create standardized PSA and Rep and Warranty agreements for various collateral asset classes. But the efforts have been quite slow and are amazingly inadequate.8 The industry efforts have been led by sell-side dominated industry trade groups consisting of dealers, issuers, rating agencies, bond insurers, private mortgage insurers and, to a much lesser degree, investors.
While these efforts could be seen as a step in the right direction, it is clear they have forged no meaningful agreement and have offered little – if anything - by way of standards.9 Instead, legislation should direct regulators to create a single standardized Pooling and Servicing Agreement governing each collateral asset class whether the issued securities are registered or “over the counter” or “bespoke”. These agreements should be created with the best interests of the investing public, and clarity of contract, at their cores.
Why Standards Matter
Legislative and regulatory standard setters must also focus on addressing a lack of clear definitions in securitization markets. Without a common language and agreement on the meanings of fundamental concepts the value of data is diminished. Conversely, if everybody is using common language – in loan origination or securitization — then it becomes very hard to game the system. The lack of clear definitions remains a huge problem that interferes with investors’ ability to compare performance of various deals and issuers and analyze and assess the true performance of the underlying collateral.
Amazingly, three years after a crisis, there is still no single standard accounting or legal definition of either delinquency or default. The entire purpose of accounting standards and securities law is to provide a framework for comparability. Yet we still do not have a single and accepted definition for so many of the key credit measures.
Currently, the term ‘delinquency’ can be determined either on a contractual or recency-of-payment basis. Even among firms that would define it on the same basis, each servicing agreement can have different interpretations of the reporting of delinquencies. Some may report advances that a servicer makes to a pool, which could be applied to reduce stated delinquencies. But other servicers may not.
How can either issuer or investor clearly understand whether they owe a duty to the other if there is so much variability from deal to deal and are no industry standard practices? Like so many of the underlying problems in the securitization market, this “Wild West, new frontier” mentality needs to be replaced with agreement of terms and standards. When no one agrees on the definition of delinquencies and or on how they must be reported, then we get a lack of standards on the definition of defaults. This leads us back to a world in which the complexity of contract is endemic to each deal and reduces the viability of securitization markets.
The Problem of Third Party Originators
Further complicating problems in the residential mortgage securitization market is the involvement of third-party originators of mortgages who are not always directly included as party to the securitization process. For example, assume that ABC Mortgage Company originates mortgage loans and sells those to XYZ Bank, which, in turn, directly or indirectly securitizes those loans. Assume ABC made representations to XYZ that were untrue, and XYZ made those representations to the investors in the securitization. And further assume that the representation and warranty agreement between XYZ (as issuer) and the investor stated that the bank would have to buy back any misrepresented loans. If XYZ had a separate agreement with ABC that required it to buy them back, in turn, from XYZ, then a larger problem could arise if the unregulated — and possibly under-reserved and undercapitalized — ABC did not have the funds to buy them back.
Over the past several years, we have heard regulatory claims that many of the problems in mortgage markets stem from the mortgages originated by unregulated third party originators. This is an unacceptable cliché that must be replaced with clear standards. If an issuer purchases mortgages from, or sponsors securitizations by, third party originators, then certain things must happen. They must be made to warrant that the originations meet their own stated underwriting criteria. And they should be required to expressly recognize any underwriting liability for any collateral purchased from third parties that does not meet their own underwriting standards. This would result in regulated financial institutions becoming responsible for ensuring their own due diligence of third party or affiliate lenders.
Simply, we need standards that transfer credit and liquidity to investors, but place underwriting risk with the issuer for specific period of time tied to the final closing of the collateral pool (after any revolving period) and linked to resets and amortization of loan specific types of collateral. If, for example, that period is one year, issuers should set up reserves against the risk of underwriting errors and servicers and investors should understand that after a year, they will no longer have the ability to put back loans for underwriting flaws unless the flaw involves fraud and it is specifically demonstrable that the fraud was directly correlated to a default.
When a pool of first lien mortgages is created and sold into a trust, a servicer is chosen to service the loans, collect the mortgage payments and direct the cash flows to investors as defined in their agreements. While investors in different tranches to the securitization may not always have aligned interests, in light of the significant numbers of mortgages today that have negative equity (close to 50%), most of the remaining holders would be willing to write down the principle balance of the loan if they would result in reperformance of collateral. For example, assume a 20% reduction in the principal balance of a mortgage would result in a borrower becoming willing and able to make payments and become current again, on a sustainable basis. This 20% loss, though significant, would surely be preferable to the potential 60% loss investors could experience upon default and a subsequent foreclosure.
Unfortunately, due to an ill-defined legal relationship between service and investor, along with a large and common conflict of interest between the servicer and the parent companies that own most of the servicers, many servicers would not prefer this “less is better than nothing” approach. Many of the servicers are owned by the largest banks –- banks that often hold the second liens or home equity lines on the underwater houses. Remember, the second lien is, by definition, subordinated to the first lien. So if the servicer wrote down the principal on the first lien, it would, where the mortgagee is in a significant negative equity position, completely wipe out the value of the second lien and cause the bank to experience a total loss on that loan.
Because of the lack of a fiduciary obligation to the first lien holder, the servicers are often motivated to protect their firm’s second lien positions, rather than the first lien holders’. And because of the way the servicing agreements are written, servicers are often able to justify their inaction by hiding behind the disparate obligations they owe to investors in different tranches. Alternatively, they are able to do so by using a “net present value test” that is based on projections of unknowable future scenarios. As a result, both investors and the troubled borrower are held hostage to servicing practices that seek to protect often under-reserved banks rather than act on their expected obligation to investors in the mortgage pool. New rules in securitization should clearly define the servicer as a owing a fiduciary duty to the investor in securitized pools. Or perhaps more effectively, it should specifically prohibit financial entities from owning servicing where the servicing results in a conflict.
Securitization has shifted significant funding for many asset classes away from bank balance sheets and into the hands of capital markets participants. With appropriate standards, securitization would more efficiently fund markets and cause a less volatile and closer convergence between the pricing and value of assets in support of economic activity and productive growth. This change is the reason that we must now restart the securitization markets.
If they are not functioning as an alternative to portfolio lending, where economically less expensive, then there is no way to finance an economy that has previously been funded by the global capital flows through capital markets. Financial institutions do not have the balance sheet capacity to directly support all or a substantial proportion of the credit previously provided by the capital markets. A failure to foster this intermediation external to the depository system will risk an ongoing shrinkage of market funding for economic activity. This, in turn, will precipitate greater bouts of deflation as access to credit remains highly constrained.
Securitization has served as a critical tool of intermediation and must be revived or replaced with a more viable tool if we are to maintain the lending capacity required by our modern economy. Functioning securitization markets, cured of information asymmetry and misaligned incentives, could help to stabilize the present situation.
While there are other areas of further changes that may be worth considering — including structuring standards and pricing and valuation enhancements which could ultimately allow securitization tranches to trade on exchanges and behave similarly to closed end funds — the recommendations in this paper are fundamental precursors to financial intermediation and must be implemented as standards for securitization or it alternatives (such as the immature covered bond market).
Joshua Rosner is a Partner at independent research consultancy Graham Fisher & Co and advises regulators and institutional investors on housing and mortgage finance issues. Previously he was the Managing Director of financial services research for Medley Global Advisors and was an Executive Vice President at CIBC World Markets. Mr. Rosner was among the first analysts to identify operational and accounting problems at the Government Sponsored Enterprises and one of the earliest in identifying the peak in the housing market, the likelihood of contagion in credit markets and the weaknesses in the credit rating agencies CDO assumptions.
The views expressed in this paper are those of the author and do not necessarily reflect the positions of the Roosevelt Institute, its officers, or its directors.