All industries are in the pursuit of high-quality products and services, and the mortgage industry is no exception. Lenders strive to produce quality loans and avoid repurchase risk, and purchasers want to make sure their purchase is within guidelines and compliant with regulations.
Industries such as manufacturing offer an abundance of best practices, tools, and techniques to produce products with no defects. They encourage quality in their products through imposing their quality standards on their suppliers and ensuring that each manufacturing step passes only defect-free components to the next step in the process. Quality is built into the product throughout the manufacturing process, and customers, as well as other stakeholders, can rely upon a very high level of quality.
Translating this concept to lending sounds simple enough, right? Lenders should be able to just adopt these best practices and meet the expectations of investors, GSEs, and regulators. Well, despite a significant investment in loan quality and compliance, mortgage production quality still has room for improvement. For instance, Fannie Mae reports that examination of post purchase reviews of 2021 acquisitions showed a notable increase in eligibility violations in key defect categories. The most frequent areas for increased defects include self-employment income, employment, and undisclosed debt.
To be fair, achieving stellar loan quality is challenging. The mortgage process is inherently complex, with information coming from many directions, many stakeholders across multiple companies involved in the manufacturing of a loan, and key loan information potentially changing throughout the loan life cycle.
Key sources of information on loan quality include both prefunding QC activities performed by lenders and post fund reviews conducted by investors. These review cycles provide information that point to specific defects, allowing it to be categorized according to a loan defect taxonomy. If warranted, repurchase requests may be subsequently issued. Corrective actions for reducing defects typically involve better training of staff and increased levels of pre-fund quality control. However, these types of correction actions are time-consuming and expensive as they rely on improving the performance of a predominately manual approach. Taking these corrective steps can require a substantial investment in training, monitoring, and incentivizing.
As Mr. Wonderful on Shark Tank says, “There’s got to be a better way!”
Building loan quality in from the start is the low-cost option and the path to stellar loan quality. Achieving unbeatable loan quality requires changing the way loans are manufactured to include loan quality considerations from the start and at every stage. While many lenders are currently implementing various forms of automation, we believe that lenders should include loan quality as part of their journey in implementing their digital mortgage strategy. The changes do not need to be revolutionary or “big bang” type of initiatives – in fact very few lenders can rebuild their loan factory all at once, and that may not even be advisable!
The “better way” to dramatically enhance loan quality involves organizing improvement efforts around two key principles. The first is automation: it’s very hard to achieve a high level of quality while operating with a largely labor-intensive approach. Most studies place the labor component of the cost to produce a loan at over 75%. Despite staff doing their best, a manual approach has an inherent defect rate higher than that of an automated approach. Lenders would be well served to ensure that their automation efforts target their high-risk areas for loan quality.
The second key principle is that the model of “inspecting quality” after the product is produced (i.e. post fund review) or when the product is almost done and expectations are set (i.e. prefunding QC) should be flipped where QC happens as part of the manufacturing process instead of at the end. Having QC as one of the last steps in the process makes the whole approach reactive instead of proactive, and the correction of defects is even more costly since most of the labor has already been invested and the loan is about to be closed.
Instead, here is a framework that we have found helpful in the past for building loan quality into the manufacturing process. Not only that, but they can also be used as guidance for streamlining origination processes for better efficiency.
- Simplify – As a part of the current performance improvement efforts, seek opportunities to simplify processes and eliminate the possibilities for errors to occur by the nature of the new process. Situations like determining which file in the file management system is the “live” copy, can be improved through streamlining processes and promoting loan quality via a new process that provides for a source of truth for documents.
- Automate – The traditional approach of “checkers checking the checkers” is expensive since the manual labor involved is not perfect and cannot be expected to perform flawlessly. Traditional checklists that are applied manually do not scale and provide little data that can be used to drive improvements and AI initiatives. Validation should be automated wherever possible and included at every step of the loan life cycle.
- Integrate – Each step in the manufacturing process needs to be coordinated with the prior steps. Lender personnel should be able to trust that the information they are dealing with is accurate and free of defects. This trust should exist for prior steps that are both manual and automated, and when in place, you should find that entire activities (such as “re-checking” data and documents) can be eliminated.
This framework can help shift the loan manufacturing process from one of inspecting quality at the end of the process to one of building quality from the start.
Technology advancements are making a modern loan factory much more possible and cost-effective, and they don’t have to be large high-risk projects to provide results. More tactical improvements focused on a single use-case or milestone in the loan manufacturing process can have significant ROI. Below are a few examples of recent tech driven improvements that are promoting better loan quality:
- Software Bots that keep the file management documents organized: Traditionally keeping file management systems organized and making active documents easily identifiable is challenging given the shared responsibility for maintaining the filing system. Bots are coming to the aid to help administer the documents automatically.
- Machine Learning enabled document validation: Validation of issues that span loan file and multiple documents can be addressed with ML (for example, verifying correct application of self-employment information across documents)
- Confirmation of accurate MI pricing: Key information on the loan file may change that requires an MI pricing change; technology that monitors for these changes and prompts the needed updates is incorporated into the normal workflow to prevent stale MI pricing.
This new enabling technology relies on new data that is largely not available in many current mortgage tech stacks. Today’s manually driven validation activities, including execution of traditional checklists, don’t produce the data needed to drive AI based processes and other diagnostic activities.
Loan quality programs need to consider adding to and extending the data that comes from traditional activities by adding data from automated validation built into the loan manufacturing process. This data should also be available to lenders’ team members so that all departments would have visibility to loan quality across the manufacturing life cycle. This information is tremendously valuable to identify not only loan quality but efficiency improvement opportunities that go unnoticed without such data.
The Path to Unbeatable Loan Quality
What is the cost of poor loan quality? We have all seen emergency drills and delayed closings as defective loans are identified only as they enter the closing department – they then require immediate attention by an array of people. We have all seen the costs of monetary cures, compliance related fines and non-saleable loans. When the total cost of defects is compared to the cost of implementing a quality loan manufacturing process based on building quality in at each step, then loan quality pays for itself. Automated and Integrated quality controls are the path to unbeatable loan quality. Along this path back-end QC processes matter less since the QC activities are imbedded throughout the manufacturing process. This imbedded approach also enables lenders to add additional controls more easily, and more quickly detect new risks. Experience is showing that building loan quality in from the start with automation is the low-cost option and the path to stellar loan quality.
Kathy Mantych is a Fintech pioneer in complex technology environments with a comprehensive background spanning 35 years. Her accomplishments and accolades include business and strategic development and effective leadership to drive revenue with an enterprise-wide, customer centric sales mentality. To be successful it takes hard work and committed focus on balancing all channels of the mortgage banking industry with experience and expertise. Kathy has had proven success year over year with multiple recognition moments and was most recently awarded 2023 Most Powerful Women in Fintech from Progress In Lending. She is currently Senior Director of Business Development at Fintech, Silverwork Solutions, LLC in Chicago promoting cutting edge technology in Robotic Process Automation with Bot technology and Automated Intelligence.