Data and HMDA Compliance – A Dangerous Intersection?

By Fiona McNeill & Matt Schriner

As banking regulations increase, useful, clean, searchable and secure data becomes a critical necessity for compliance officers. It can mean the difference between fines or successful mergers, public humiliation or public relations victories. Unfortunately, many compliance officers view data as a necessary evil – something they end up defending rather than something that can head off a crisis and even help build market credibility and profitability for their institutions.

The regulatory environment is in a constant state of change. But one thing is certain – regulators are ratcheting up the stakes. As a result, banks and mortgage companies must gather, maintain and interpret data – especially that which is covered by the Home Mortgage Disclosure Act (HMDA), designed to root out predatory and discriminatory lending.

In 2003, regulators increased their demands on compliance officers by requiring the implementation of new HMDA data elements, including the collection and publication of lending rates. As a result, lending data from 2005 can be easily accessed and viewed online by consumers and competitors. Consumer advocates continue to push for additions to the HMDA regulations that will provide more information on borrower risk and the tracking of high-cost loans. States, too, are requiring more paperwork to regulate broker transactions and redefine local anti-predatory lending laws.

Meanwhile, most compliance officers consider the costs necessary to meet these requirements a burden. But in fact, managing data doesn’t need to be a money loser. With the right approach, proper data management coupled with analytics can generate a better bottom line for the financial institutions. In contrast, the costs of mismanaging data and forgoing analysis can be steep.

The negative effects of bad data
In 2006, one of the nation’s largest sub-prime lenders agreed to a multimillion-dollar abusive lending practice settlement with attorneys general from 49 states. By that time, the company had already shelled out $100 million in lawsuit settlements, and it continues to face dozens of suits.

That company joins a long line of banks and mortgage companies that have been fined or forced to settle when faced with predatory lending and fair lending charges following investigations sparked by HMDA filings. Beyond the fines, companies in violation of fair lending laws also receive negative ratings, which can affect customer opinions and regulatory approvals for new services, acquisitions and mergers.

The most recent lender to get fined has admitted no wrongdoing. However, one thing is apparent – the company’s executives had little if any help identifying the rogue sales staff and loan processors who reportedly inflated appraisals, mortgage rates and applicant incomes to generate business. The typical process for submitting HMDA data involves gathering it up for the entire year and frantically tabulating it in time for the March 1 deadline. Historically, executives have not had scorecards that can give them a monthly, weekly or daily view of what is going on with their business. The result for the company, along with the fines, may be a massive reorganization including the closure of branches and loss of jobs for thousands of workers.

Start with clean data
Even if banks and mortgage companies carefully review their sub-prime lending in detail and keep a close eye out for evidence of potential predatory or discriminatory practices, few employ rigorous statistical tools capable of rooting out rogue processors or providing solid data to counter critics.

Sub-prime lenders argue that they offer loans to people who wouldn’t otherwise qualify because of poor credit history. Critics argue that not all of these borrowers are high credit risks. Re-examining the factors that went into establishing the price or loan-underwriting conditions, compliance officers can now conduct automated match-pair testing – surfacing the most glaring differences to focus effort for further investigation. This technique readily identifies similarly situated protected class borrowers and can help counter criticism that the bank has shown bias. Such testing can also uncover any potential problems prior to bank examiners’ manual file reviews.

Using business intelligence software could allow banks to drill down into their data to look at specific lending practices by branch, by product or by geography to get a heads-up on trends suggesting disparate treatment or the need to educate staff better. A bank could also compare its HMDA results to those of its competitors. It takes more than a packaged tool to provide this type of service to financial institutions – even if that tool is accessible to more than one person and is housed on a server. Financial institutions need clean, auditable data that is centralized to provide the same, timely facts to everyone in the organization – whether they are cleansing it for regulatory submission, reporting from it for business understanding or analyzing it for fair lending.

HMDA data is often gathered on desktop computers. Banks rarely employ data quality scrubbing software that checks for inaccuracies or even use exception-based flagging of edit errors. Instead, when banks discover that their data is inaccurate – or, even worse, when government examiners discover it first – the typical response is to bring in high-priced consultants charging hundreds of thousands, even millions, of dollars to straighten out the mess on a one-time basis. Those charges do not include the expense of government fines and increased insurance costs. So isn’t it better to start with clean data?

It’s not just about complying; it’s about analysis and explanation
For most compliance officers, a call from the company’s public relations department is not good news. Notices from the legal department are even worse. Banks and financial institutions often resort to defensive posturing when the HMDA data isn’t so good. All the positive spin in the world isn’t going to make that problem go away, or keep the regulators, attorneys general and plaintiffs’ attorneys at bay.

What will keep them away is strong analytics. If a bank can analyze the data, it can root out problems before they become part of the public filings. With the help of visualization, internal officials may readily learn what they need to do, while making the processes transparent to external customers. Being able to analyze the characteristics of mortgage customers, the conditions around the loan (such as the property information, the channel, etc.) and their final loan costs show everyone what is really happening. Compliance officers can make recommendations rather then react defensively to poor data outcomes if they are armed with a cohesive, enterprise data strategy that is robust enough to answer any question, scalable enough to grow with the business and flexible enough to accommodate the ongoing dynamics of regulatory changes.

*A longer version of this story originally appeared in the Jan./Feb. 2007 issue of ABA Bank Compliance magazine (Vol 28, No.1).

Bio: Fiona McNeill is a Principal Consultant for SAS Fair Banking working with customers on technology needs and defining corporate analytic product strategy.
Matt Schriner joined SAS in February of 2006 as Business Director for SAS Fair Banking. Prior to SAS, he served as Managing Director for a national risk management consulting group and was a compliance examiner with the Federal Reserve.