Monday, August 5, 2013

Brace for Impact: GSEs Announce Uniform Closing Dataset (UCD) for New Closing Disclosure

For over a year, title agents and loan originators have been closely following efforts by the Consumer Financial Protection Bureau (CFPB) to create new rules and disclosure forms for real estate settlements. Under the proposed rule, expected to be finalized this fall, a new Closing Disclosure form will replace the existing HUD1 settlement statement, while another new form, the Loan Estimate, will replace the Truth in Lending (TIL) and Good Faith Estimate (GFE) forms.

The new Closing Disclosure, in particular, will require significant changes to software and office procedures. Gone are the familiar line numbers of previous HUD1 forms - replaced by a new format that alphabetically organizes loan costs into new groups, such as "Origination Charges", "Services Borrower Did Not Shop For", and "Services Borrower Did Shop For". (This of course, assumes that the final disclosures will resemble the earlier drafts - and all indications are that they will.)

New rules also require that the Closing Disclosure be provided to the consumer three days before closing, with any significant changes to costs triggering a new waiting period. This rule will require loan originators and settlement agents to share information much more closely in order to be compliant.

Although the draft CFPB disclosure rules put ultimate responsibility for the accuracy of the Closing Disclosure on the lender, most industry insiders fully expect title and settlement agents to maintain their role in preparing this statement. In order to stay compliant with the three-day rule, lenders and settlement providers, each of whom provide some of the cost information for settlement statements, will need to make sure their software systems can exchange closing costs electronically.

In an effort to standardize the underlying data required by the new disclosure, Fannie Mae and Freddie Mac have jointly created a new standard dataset called the Uniform Closing Dataset, or UCD. This dataset will be a new component in the Uniform Mortgage Data Program (UMDP), which is an ongoing endeavor by the GSEs, under the direction of the Federal Housing Finance Agency (FHFA), to improve the quality and accuracy of loan data.

Similar to other UMDP standards, such as ULDD (Uniform Loan Delivery Dataset) and UAD (Uniform Appraisal Dataset) which are already widely used, the UCD will be based upon XML standards produced by the Mortgage Industry Standards Maintenance Organization (MISMO).

Although full details on the UCD will not be released until after the CFPB releases its final disclosure rules later this year, this is a very significant development for several reasons and offers organizations and vendors a real opportunity to reduce implementation costs and risks associated with the new disclosures.

For software providers who must update systems to support the new Closing Disclosure, the UCD will take much of the guesswork out of determining which pieces of data go into which spots on the form. The MISMO Origination Workgroup, with feedback from the CFPB, has documented how to populate the new disclosures using the MISMO data points which underlie the upcoming UCD.

In addition to providing a blueprint to populating the disclosures, the UCD will also serve as a standard data format for settlement software and loan origination systems (LOS) to exchange information about loan costs. By embracing these standards, software providers have the opportunity not only to comply more easily with the CFPB's three-day rule, but also to increase the pool of trading partners with which they can easily integrate.

The most important thing to note about the UCD is that, like the ULDD and UAD before it, the dataset is likely to become a required format for all loans acquired by the GSEs. This is not something settlement service providers are going to be able to ignore.

With these new requirements coming, title agents and lenders should take time now to verify that their respective systems can accept electronic information from the other. Lenders, in particular, will likely demand this capability from settlement providers in order to ensure compliance with the three-day rule. Rekeying of data from one system to the other is not going to be an acceptable practice under the new rules.

Rather than waiting and "bracing for impact" later on, settlement providers and lenders would be well-advised to stay informed as more information comes out about the UCD and incorporate it into their plans to implement the new disclosures. Hopefully, adhering to and using the guidance inherent in the UCD will make the transition easier and less costly rather than simply being an added burden.

Sunday, July 28, 2013

To Minimize Integration Costs, Make Your Application a "Native Speaker" of MISMO

If you've ever been to a foreign country where the predominant language is unfamiliar to you and English is not widely understood, you probably experienced some frustration in trying to communicate. Doing something that is ordinarily simple, such as asking for directions, becomes a challenge when you must rely on your own limited knowledge of the foreign language to formulate your question, and then struggle to understand a response.

Ideally, you would like to find a local who also speaks English in order to have an easier time, but even the task of trying to locate that person takes time and effort. Imagine how much easier and enjoyable your trip would be if you could speak the foreign language fluently? No one speaks a given language more easily that a true native speaker - someone who learned the language during their formative years.

Software applications often struggle with similar translation issues when they must communicate with other systems. Commonly used integration methodologies such as REST or SOAP web services only address the mechanics of communicating, not the semantics. This would be akin to foreign speakers agreeing to communicate by forming distinct sounds with their vocal cords, lips, and tongues, but saying nothing about words and phrases.

For two applications to communicate, they must also agree upon semantics. One application must be able to send data (formulate a sentence) or request some data (ask a question) from the other application using terms (words) understood by both parties. If the two applications use different terminology or have different expectations on how to communicate, then extra effort development effort is typically required to build some sort of translator or adaptor layer. This adds to both the cost and the length of time required to make two applications talk to each other.

Ideally both applications would have been designed from the beginning to speak a common language, meaning that the underlying data model, service calls, and protocols for "talking" would be similar and no translation effort would be needed. The two applications could just converse - as easily as you might have a conversation with a neighbor.

For the mortgage industry and affiliated service providers, such as title insurance underwriters and appraisers, this common language exists in the form of the MISMO reference model. MISMO defines a common set of industry terms (words) as well as XML standards for common data exchange scenarios (questions and sentences).

An application that is a true "native speaker" of MISMO stores its data internally in a form similar to the MISMO reference model. In other words, the table names and relational structure mirrors the XML structure of MISMO. The column names are also aligned with common terminology from what MISMO calls its Logical Data Dictionary.

When coupled with a service layer that uses MISMO standards for forming the "sentences" of requests and responses to other systems, an application is ready for easy, low-cost integrations with other compliant systems. No translator required.

Since MISMO is rapidly become the de facto standard for mortgage and real estate transaction data, building new applications as "native speakers" will enable them to interact and adapt both now and in the future.

Tuesday, June 25, 2013

Don't Call it Title Insurance - Call it a Warranty

I recently read about yet another case of title insurance claim payouts being unfairly compared to other types of insurance, such as property and casualty. During a recent congressional house hearing on Qualified Mortgage, the president of the Center for Responsible Lending made a comment lamenting that, for every dollar spent on title insurance, 75 cents goes to commission and 10 cents to claims, while other forms of insurance pay 80 to 90 cents on the dollar for claims. These statistics are very misleading, and labeling title policies as "insurance" rather than a "warranty" is a big reason behind the misinformation.

When a lender or an individual buys a title policy, what they are really paying for is the effort that goes into ensuring that a piece of real estate has clean title, rather than insuring against title issues that might crop up in the future. This effort includes property searches for things such as outstanding liens that might cloud title, as well as work that is done to actively cure adverse conditions. A common example of curative work is getting lien releases from prior payoffs properly recorded in the county recorder's office.

A claim against a title policy is an outcome that is desirable to no one - neither homeowner, lender, nor underwriter. Everyone involved would prefer that the stress, expense, and aggravation of a claim be avoided by the work and due diligence that is done before a policy is issued.

A title policy is really a guarantee on the quality of the work that went into searching, documenting, and curing title defects. This is more similar to a warranty on a new car than to traditional insurance, such as a homeowner's policy. When someone purchases a new car, they are issued a warranty that guarantees, for a period of time, the quality of materials and workmanship that went into making the car. No car owner ever wishes for the inconvenience of unexpected car trouble, whether the repairs are covered under warranty or not. The ideal claims rate for a car warranty is zero. Title insurance should ideally have a very low claims rate as well.

Perhaps the best way to stop misinformation and unfair comparisons is by emphasizing to consumers and regulators that title premiums are mostly paying for the work that goes into ensuring clean title rather than insuring against the uncommon scenario of future claims. Using the term "warranty" rather than insurance might be one way to make this distinction clearer.