Putting XBRL to Work in the Mortgage Market

Unless you have some combination of a very strong credit rating, a high income-to-debt payment ratio and a relatively low loan-to-value ratio, it’s not especially easy to refinance a mortgage these days. That’s a shame, because there are plenty of people who have stayed current in meeting their credit obligations and whose mortgages are comfortably below current market value who could benefit from today’s record low interest rates. One major reason they can’t refinance is the collapse of non-agency mortgage-backed securities (MBS) – that is, those not backed by government agencies such Federal National Mortgage Association, or Fannie Mae – in the wake of the 2008 financial crisis. The crisis, in turn, was caused largely by the collapse in value of mortgage-backed securities. To be sure, a significant portion of the drop in the issuance of non-agency paper is the lack of demand these days for the risky and even fraudulent sub-prime mortgages that were a root cause of the financial collapse. Yet there would be a bigger market for MBSes (and therefore more money available for refinancing) and less need for U.S. government guarantees if there were greater transparency in the quality of the underlying assets of mortgage-backed securities. Technology exists today that would address the transparency issue relatively easily and inexpensively. In particular, eXtensible Business Reporting Language (XBRL) can provide an efficient and relatively inexpensive means of collecting needed information from a large number of disparate parties without requiring them to standardize or modify their systems.

In the post-financial-crisis environment, many have questioned the inherent soundness of MBSes. There’s nothing fundamentally wrong with securitizing any debt in an asset-backed security (ABS), a broader category of tradable debt that also includes such things as car loans and credit card obligations. Indeed, asset-backed securities were invented to enable their issuers to transform a group of smaller, illiquid investments (such as individual home mortgages) into a larger, tradable and more diversified investment. Historically, the process of securitization has improved the efficiency of the credit markets, in part by making assets available to a broader set of investors and enabling those holding the securities to diversify their geographic exposure and other risk characteristics.

However, as they are currently structured, mortgage-backed securities lack transparency into the current state of the underlying assets. Until the credit market collapse in 2008 this was of little concern. Some of this nonchalance was the result of assumptions that developed before the age of securitization and that didn’t change with the times. Back then, the company that originated a mortgage loan or other collateralized debt usually held onto the obligation. Since it would be “holding the bag,” it took pains to ensure the creditworthiness of each borrower and the soundness of the collateral. Moreover, because it also serviced the obligation and had a general idea of market trends that affected the value of the underlying assets and the ability of the borrowers to make their payments, the owner of the mortgage also was able to monitor the ongoing soundness of the credit.

The lack of transparency into the current state of the underlying loans and their collateral is a product of securitization. When loans are bundled and resold, the constituent pieces of the mortgage process (origination, servicing and owning the credit) are split up. This division has rendered the up-front quality control in the origination process all the more important because it is the first line in ensuring credit quality. The attestations made in this phase also are a key driver of the initial credit rating of the obligation. The arrangement worked well at first, but in the early- to mid-2000s complacency (or worse) in vetting the origination process, abetted by demand for higher yields, compromised the process (to put it mildly) and was a major factor leading to the crisis. As well, easy credit in marginal housing markets drove up prices to unsustainably high levels. When loans became unavailable, housing prices in these markets collapsed, and it became difficult and time-consuming for owners of MBSes to know to what extent the underlying mortgages backing a specific certificate exceeded the value of the property. This systemic point of failure created uncertainty that froze credit markets generally.

Since the crisis, there have been calls for reforms that, in effect, attempt to turn back the clock. However, rather than trying to address this year’s problems with last year’s solutions, it’s possible to take today’s disintegrated mortgage finance business model (the separation of origination, servicing and ownership of the debt) and make it work at least as effectively as the integrated approach. Transparency is the key.

Transparency in mortgage-backed securities means that rather than relying on prospective credit judgments initially and for years after, each MBS certificate can be assessed at any time to determine its creditworthiness. A prospective buyer should be able to know the creditworthiness of an individual security and its constituent parts based on up-to-date information at the time of purchase. (I take up-to-date to mean data that is no more than six weeks old.) For example, it is possible to determine the current loan-to-value (LTV) ratio of a specific mortgage by looking up the estimated price of the house on Zillow, a real estate website. Since Zillow bases its price estimates on recent comparable transactions in a specific area, it should be reasonably accurate and not susceptible to fraud. The current creditworthiness of a loan’s borrower (and co-signers) can be checked quickly with credit bureaus without compromising an individual’s privacy. However, the main obstacle to achieving the needed level of transparency stems from the inability to get information from the mortgage loan servicers – companies that collect loan payments and pass them along to the owner of the mortgage. Payment history is a key element in assessing the riskiness of a mortgage, but servicers have been unwilling to provide this or other information, either at all or without compensation.

To address this issue, the securities industry and institutional investors must define the data required of the servicers, ensure mortgage contracts include a provision requiring the servicer to provide the information to a data aggregator (which might be the Mortgage Bankers Association), arrange a compensation mechanism for the servicers for providing the data, agree upon a method for collecting money to pay for the information-gathering process (both the servicers and other data providers) and create an efficient method of transmitting and storing the servicing data. The last-named is likely the easiest to do.

The fact that the myriad mortgage servicing companies use different systems, different data models and different operating systems is not a barrier to efficiently collecting data. Once the individual data elements required from the servicers are defined, the servicers need only generate periodic (say, weekly or monthly) electronic reports using XBRL to define and tag individual data elements so the report can be consumed by any data store. Automating the tagging process and validating the data is a straightforward, well-established and inexpensive process similar to other XBRL-tagged data transmission efforts worldwide, such as reports to banks regulators in the United States, Japan, France and other countries.

Why was this approach not implemented when asset-backed securities first appeared in the 1980s? Simply because it was not feasible until recently. Decades ago, when ABSes and MBSes first appeared, the expense of tracking the state of each and every piece of the underlying debt was prohibitively high. Today, this is no longer the case. The costs of computer memory and processing speed are fractions of what they were, the Internet is an ultra-low-cost common network and XBRL makes it possible to easily and inexpensively provide the necessary data from a large number of disparate and incompatible systems.

It’s possible that over the next year or so an arrangement with mortgage servicers can be worked to make the relevant data available. If that happens it should become feasible for prospective owners and investors in mortgage-backed securities to quickly assess the credit quality of individual securities. Once this innovation is in place, I expect additional credit will become available to the housing market.


Robert Kugel – SVP Research

2 thoughts on “Putting XBRL to Work in the Mortgage Market

  1. This is a great article which shows how XBRL can be applied. But what was left out of the article is that the reasons XBRL might work for the mortgage market are the same reasons XBRL is more generally applicable to other similar problems as well:

    See a list of these reasons and the general features XBRL brings to the table on my blog post here:


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s