Enhancing Servicing Insight Through Big-Data Analytics
Big data will hit its stride when tailored to vertical markets and their specific needs. Mortgage servicing is an excellent example.
Mortgage servicers face significant challenges. Declining revenue resulting from a precipitous drop in total balances combined with increased operational overhead driven by new regulatory requirements and customer demands have created a stark reality—remaining profitable, sustainable and scalable demands innovation. Fortuitously, mortgage servicers have access to a particularly valuable and largely unleveraged asset—data.
Mortgage servicing is a hyper-competitive, low margin, high transaction volume business, and it serves an industry that has shrunk over 12% since 2008 (from $11.3 trillion to $9.9 trillion). Successful servicers—survivors and thrivers—will be those that intelligently apply sophisticated analytics and ask smart, insightful questions when presented with mountains of unrefined data—big data. Better-informed decisions facilitated by new perspectives regarding mortgagor behavior will result in enhanced risk mitigation, improved loan quality, higher per transaction margin, and increased profitability.
Enhanced data analysis will deliver a deeper understanding of why mortgagors act as they do. Mortgagor behavior must be better and more holistically understood, as it drives servicer success. Consider the following partial list of key elements that directly impact success:
These factors, among others, drive mortgagors’ propensity to: refinance; default, return to your firm for their next mortgage, accept other products your firm may offer; and contact your customer service area.
Yet, servicers tend to focus on interest rate arbitrage and loan-to-value ratios as the sole drivers for anticipating mortgagor prepayments and defaults; the two largest drivers of value. Other data that have been shown to strongly correlate with mortgagor behavior have been ignored, such as FICO, inflation adjusted LTV, method of payment (coupon book versus ACH), the existence of second mortgages and other debt, payment history, current principal balance, marital status, local migration patterns, unemployment rates, health and wellness, and myriad other demographic factors.
Similarly, servicers tend to lack applied data analytics sophistication when considering customer recapture rates, customer service demand, and cross-sell opportunities. For example, many servicers report strong recapture rates. Their loan officers’ compensation structure however incents high recapture rates only when rates are high. As rates fall and production grows, loan officers are disproportionately incented to close on new borrowers and ignore existing customers. Loyal and profitable cohorts of the servicing base are thus pushed, inadvertently, to the competition. More sophisticated, institutionalized data analysis will render better compensation structures and more profitable business outcomes.
Big data can further benefit servicers by providing access to valuable data not currently part of any calculus. Because we don't know what we don't know, new capabilities to access and process traditional and new structured and unstructured data in real-time offers servicers new perspectives that provide true competitive advantage—innovation through deeper insight. Many external data sources can be infused into existing data to further tease the signal from the noise, such as:
- Census data
- Social media data
- Origination data
- Historical servicing data
- Housing price indices
- Economic data from a variety of sources
- Bank Call Reports
- Trade association research
Success demands servicers understand properties and their locations, loan products, and most importantly mortgagors as individuals. In a business predicated on narrow margins, high transaction volume, and risk management, mortgagor behavior is not only relevant it is a critical risk variable commanding rigorous contemplation.
While embarking on such an analytics initiative appears daunting, there are mortgage technology services firms that leverage cost-effective computing and communication resources, advanced integration techniques, subject matter expertise, along with external data sources to bring big data principles and technologies and sophisticated analytics to mortgage servicers of all sizes.
Basic demographics, personal debt, payment tendencies, and the relationships between payment status and chance for default are commonly available. However, what behavioral markers are missing? If, for example, a servicer understands why a mortgagor suddenly ceases timely payments, would such awareness provide a deeper understanding of the individual case? Can higher fidelity risk assessment be far behind? The ability to optimize opportunities and minimize overall risk using data and analysis previously inaccessible and do it quickly is available now. All mortgage servicers can actively participate in the big data "movement" and develop more valuable insight. Newly available data can and should be leveraged to gain a competitive advantage. Seeing into the future (e.g., predicting individual behavior) with accuracy is, after all, the gold standard.
Tom Healy, CEO, Level 1 Loans, also contributed to this blog.