Investment firms are also increasingly unwilling to trust the work of others. This is coupled with the need for more granular information down to the borrower level. These have driven market participants to build their own models using more detailed data from various sources and are closely tied to increased transparency in transactions.
"Investors want to analyze the data themselves, and we provide a platform that allows them to do that," said Greg Munves, vice president at 1010data. "When dealing with large data sets like loan and borrower-level data, you need a platform that allows easy access and flexible analysis of all available information. This capability is crucial to the construction of more accurate, proprietary models. Of course third-party models are still important as benchmarks, and they also need to be based on this level of detailed information."
Access to loan-level data has always been available to the market, but the prodigious amount of work required to analyze such detailed information limited its use during the "good years."
When the market became more competitive, such analysis was no longer optional and, as the available information expands, the industry will have even more data with which to contend. "Analyzing the universe of non-agency deals is a relatively large data problem unto itself, and now we've brought borrower credit data into the picture, exacerbating the problem," Munves said.
Munves added that, by using updated credit data, investors can significantly improve the accuracy of relative value analysis.
"If you are comparing multiple pools of loans that look very similar based on the loan-level data, and you now have twenty new pieces of up-to-date information about the borrower, identifying relative value between the pools becomes a more meaningful exercise that can lead to better decisions about what you want to do with the pool," he said.
Standard & Poor's and Experian Capital Markets recently announced that they will provide consumer credit scoring data for the individual mortgages in RMBS portfolios. The consumer credit scoring data will be integrated into S&P's securitized loan data feeds.
This is the next step in S&P's ongoing effort to bring more detailed information on the underlying collateral in MBS to investors, governments and regulators.
The alliance will provide investors worldwide with more detailed information on the underlying loans in U.S. MBS by combining Experian's consumer credit data and analytics and S&P's loan-level data products. In the first step of the partnership, Experian Capital Markets will connect consumer credit information and attributes to S&P's U.S. RMBS Edition loan-level data feed product.
"The goal of our collaboration is to provide investors with the transparency needed to value structured finance products and to make more informed buy and sell decisions," said Ethan Klemperer, senior vice president and general manager at Experian. "Our partnership with Standard & Poor's is a critical step in improving market efficiencies needed to restore liquidity and investor confidence."
In February, Equifax launched ABS Credit Risk InsightTM, a solution for the MBS market that enables investors to link mortgage loan-level data on the entire universe of non-agency mortgage securities to up-to-date borrower credit data.
ABS Credit Risk Insight gives investors data on leading indicators of mortgage default such as updated credit scores, balances and utilization, delinquencies and defaults, monthly payments, credit inquiries, length of credit history, owner occupancy and refinancing activity. With data updated bimonthly, Equifax's solution allows investors to better predict mortgage performance by adding up-to-date borrower risk scores and credit data to their models at the loan level.
For investors using this solution for trading and modeling, they can better predict loan delinquency, default and prepayment, identify current healthy deals and monitor changes in collateral health.
Equifax recently published research that examined the impact of silent seconds on CLTVs and debt load among current Alt-A and prime borrowers by analyzing borrower credit information, Federal Housing Finance Agency home price data and loan-level data. The analysis demonstrated how leading indicators such as the prevalence of second liens can be highly predictive of mortgage default and loan performance.
"This data used to be unavailable to investors once mortgages were securitized, but we have overcome those barriers and can provide fresh credit data for all non-agency mortgages," said Steve Albert, vice president of capital markets at Equifax. "These insights on consumer credit behavior and borrower debt loads provide investors with an early warning indicator of borrowers who are likely to get into trouble paying securitized mortgages and home equity loans."
Pricing Transparency
At the Information Management Network's ABS East 2009 conference held last October in Miami, there was a lot of discussion about ABS pricing transparency. Investors have been digging deeper into the root cause of why pricing has not been as transparent as it needs to be.
Because of the ongoing liquidity crunch and a lack of secondary market activity for many types of securitized instruments, organizations must now rely more heavily on financial models to support valuation prices.
Among the factors that impair transparency are inaccurate collateral valuation (specifically, the asset valuation of the property) and the misrepresentation of borrower quality, said Jim Yeh, chief analytics officer at Digital Risk.
However, he added that a centralized appraisal practice mandated by Fannie Mae will go a long way to address the inconsistency in valuation. At the same time lenders are tightening guidelines and demanding more document disclosures.
"The work that the [American Securitization Forum] is doing via its Project RESTART would go a long way to stabilize borrower information and subsequent information on liens and any other liabilities on loans," he said.
Had more investors been looking into this type of disclosure and had borrower information to this level been available, particularly on liens, then it's likely that at the very least the market could have seen the problems coming. "It wouldn't have prevented the crisis, but the severity of the problems that we see presently would have been less," Munves said.
However, change still is needed to detect borrower or broker misrepresentation.
Yeh believes that there still needs to be more work done on the mortgage origination side. Digital Risk - together with its peers in the market that engage in fraud detection - is offering many services that allow customers to identify fraud early.
While each has its unique features, Digital Risk's product is more data driven than what its peers offer. Digital Risk is the largest forensic loan review firm in the nation, and it is also very technologically oriented in its review process.
"Digital Risk captures more detailed review results than other review firms, and with the largest review results in the nation," Yeh said. "The score models developed by Digital Risk hone in on early fraud activities with a lot more scientific backing than other heuristic products that its peers offer."
This technology allows users to reconstruct underwriting at the point of origination and take into consideration a borrower income perspective that perhaps went untested originally. After all, part of the problem with subprime lending, particularly during the boom years, was the indiscretion in borrower behavior when reporting income, which went untested because of the lax loan underwriting standards.
"I think today investors need this information faster and at a price that is more economical, allowing a wider level of investment participation and larger opportunity to more firms," said Thomas DeLorenzo, a managing partner at MBS Data, which provides a platform for loan-level remittance data to small and mid-size buy-side investors that have been priced out of the market when trying to obtain detailed MBS deal and loan-level information.
These investors now have access to data, performance metrics and analytics on more than 6,000 non-agency MBS deals at a cost consistent with today's economy.
Investors come in various sizes, each with different data and analytical needs.
There is a need for more transparency. Price was so large that it really left some buyers priced out of the market. Now there are more players in need of information at a reasonable cost. These investors might not have had the modeling capabilities and were relying on others to do this for them.
However, the days of depending solely on data provided by institutional firms are clearly over. What the market is seeing more and more of are market participants getting their hands on data and creating their own set of analytics to see what the results are based on how they plug in that information.
The ABS industry is seeing some real interest outside the traditional investor base, with hedge funds and the three- to five-man shops beginning to incorporate some of this new information technology to create mortgage funds and REITs, according to DeLorenzo. "These guys are getting prepared for the new round of securitization, but what they believe is that they now need to have this loan-level data at the ready,"
Non-agency deals come in various forms and are all over the map, but getting down to the nitty-gritty is essential if you aim to make any yield or at the very least not lose out.
DeLorenzo also said that they are looking to expand the dataset to include agency MBS and, given the interest, they may also expand into the commercial space.
The changing space of technology is also fueling competition and commoditizing information that wasn't available to market at large before.
"There is never too much information, but it has to be the correct information if it is going to make a difference," DeLorenzo said. "The danger is that firms have to come up with a testing period to make sure what they are pulling and that there is no problem with the underlying data."
Looking Beyond the Borrower
Because of the recent troubles, the market has shifted its emphasis from borrower quality to collateral (home) valuation. Rapidly changing property values are a complicating factor. Using multiple approaches to property valuation is essential to improving the quality of valuations.
Lewtan Technologies launched its ABSNet Loan HomeValTM data feed in October. For the first time in the securitization industry, interested parties can match the securitized loans in an MBS pool to actual homes.
This is accomplished by receiving updated loan information and monthly home valuations. The loans are valued by using automated valuation models developed by Collateral Analytics. Matching loan-level data to specific properties and current home valuations enables ABSNetLoan HomeVal to determine the quality of the asset behind the loan.
In the ABSNet Loan HomeVal data feed, clients are offered four sets of data points per month, which are at the property, ZIP code and core based statistical area (CBSA) level. ABSNet Loan HomeVal fraud identification fields offer a preliminary screening for high-risk loans and identify potential fraud.
Ned Myers, chief marketing officer at Lewtan Technologies, said that the technology today is allowing access to actual property information. "Once the loans are securitized, they aren't included with addresses and borrower information due to privacy restrictions around detailed borrower information," he said. "For investors this proves a disadvantage because they do not know which properties they have and consequently make decisions based on broad estimations on what the property might be worth."
Myers said that the technology available today means that those factors can actually be modeled and matched to the property within a securitization pool.
"The challenge has always been figuring out which of these loans are backing which securitization transactions," Myers said. "We update the portfolio of loans each month, we provide the most accurate property valuations available and we match that to the securitization data from ABSNet to show how these bonds are performing."
Some industry players believe that this "data overload" factor may not help in trading bonds.
However, relative-value strategies that involve frequently trading in and out of specific securities based on the latest data could employ such technology. For these types of trades, it means knowing what the collateral is worth since one can get a better estimate of what the bond is worth. It also impacts risk management because the buy side can better understand what the credit risks are and how much capital they should be holding.
In a market where there is very little pricing transparency, determining the price of a security is virtually impossible and, in instances where the buyer or trader is armed with deep-level loan information, it could serve to create some yield on trades.
"Having more data doesn't take away the risk, but it levels the playing field for transparency of data," Myers said.
Yeh believes that investors have already experienced how bad things can really turn out in a credit crisis, and how the government is capable of influencing the securities structure.
"The future demand of MBS will place high importance on the certainty of the principal soundness and the stability of prepayment speed," he said. "Short maturity bonds with sufficient risk premium will likely be the favorite of MBS investors for years to come."
As of now, new issuance in private label securities (PLS) MBS is still rare. However, there are sufficient talks in the market of the return of PLS MBS new issuance, with a few major banks gearing up their conduit business.
It is expected that in 1Q10, there will be some visible activities in the PLS MBS new-issuance market, which will be the testing cases to see if the investors demand more disclosure or representation from the issuers.
(c) 2009 Asset Securitization Report and SourceMedia, Inc. All Rights Reserved.
http://www.structuredfinancenews.com http://www.sourcemedia.com
