In the eight years since the financial crisis, a good deal has changed in the mortgage business. The industry has seen some of the biggest changes in the area of quality control, which has moved to the forefront of most lending institutions, Regulators are looking for banks and other lenders to have comprehensive and detailed quality control plans in place. Investors are likewise focused on the quality of mortgages. [paragraph] The government-sponsored enterprises (GSEs) seem to be moving toward requiring seller/servicers to be responsible for independent quality oversight. The GSE repurchase framework was overhauled in late 2014, and in October 2015 the agencies released another revised representation and warranty framework. The GSEs appear to be looking for lenders to get it right the first time rather than face the risk of repurchases. [paragraph] Some companies are taking a broader approach and implementing overall quality-management plans, which encompass not only quality control, but quality planning, quality assurance and quality improvement.
The four components of quality management are defined as:
* Quality planning--This step identifies requirements, sets criteria, and outlines proper processes and procedures.
* Quality assurance--This helps ensure that consistent results are provided. Quality assurance helps organizations avoid defects and mistakes in the mortgage loan process.
* Quality control--The quality of products or services are reviewed. This component includes testing to identify problems or issues that need correction.
* Quality improvement--Through quality improvement, the results can be measured and improvements in processes or procedures are made.
Even with all the efforts under way, there's still progress to be made. Many companies are taking piecemeal, reactive approaches to testing, monitoring and managing loan quality. Additionally, communication between departments is lacking and the result can be a fragmented and largely ineffective quality-control program.
Servicing quality is improving
It's certainly no news flash that mortgage servicing is being scrutinized heavily by regulators. The Consumer Financial Protection Bureau (CFPB), investors, the GSEs and Department of Housing and Urban Development (HUD) all require servicers to continually monitor and audit their servicing departments.
With investors and regulators requiring increased clarity on how loans are serviced, mortgage servicers have been obligated to make drastic changes in the way they do business. Mortgage servicers have had to develop new processes, and in some cases build new servicing organizations.
Managing the quality of servicing has become extremely important in the wake of new regulations and directives over the past several years.
Chief among those is the National Mortgage Settlement (NMS) agreed upon in 2012, which created new servicing standards for the banks covered under the settlement--five of the largest servicers in the country. The NMS established nationwide reforms to servicing standards, which required better communication with borrowers, a single point of contact (SPOC), adequate staffing levels, training and appropriate standards for executing documents in foreclosure cases. Today the settlement covers seven of the nation's largest servicers.
The settlement struck a powerful blow to the top banks, which have spent significant money to comply with the terms of the settlement. But banks are realizing that certain mandates of the settlement are providing real value to their organizations. For example, the internal review group (IRG) model appears to be working and will likely continue in some form well past the settlement end date (see sidebar, "The IRG Model Is Here for the Long Term").
Perhaps the most significant consequence of the reforms to mortgage servicing standards are the organizational habits servicers have created following three years--or 12 quarters--of oversight to ensure the implementation and ongoing monitoring of the end-to-end servicing process.
In particular, the NMS and regulators have focused on servicers' unfair and deceptive interaction with borrowers. For example, in the CFPB's summer 2015 issue of Supervisory Highlights, examiners found at least one servicer that sent notices of intent to foreclose to borrowers already approved for a trial modification.
Then, in its fall 2015 issue of Supervisory Highlights, the CFPB outlined several ways that examiners found servicers had violated Regulation X, which implements the Real Estate Settlement Procedures Act (RESPA) of 1974. Effective January 2014, the CFPB amended Regulation X to modify and streamline certain servicing-related provisions. The violations reported in the fall of 2015 included:
* Upon the death of a borrower, one or more servicers lacked any policies and procedures for identifying and facilitating communication with successors in interest.
* One or more servicers sent letters to certain borrowers soliciting loss-mitigation applications when the servicer's own records showed that the borrowers were not eligible for any loss-mitigation option.
* One or more servicers evaluated the borrower only for the loss-mitigation options that a servicer representative preselected for the borrower.
* In one or more instances, a servicer's foreclosure attorney sent a foreclosure referral letter to the borrower after the borrower had already entered into a loss-mitigation agreement with the servicer.
But progress has been made. Late last year, the monitor of the Office of Mortgage Settlement Oversight, Joseph A. Smith Jr., found that six of the seven mortgage servicers subject to the terms of the settlement had achieved complete compliance with the NMS' servicing rules in the first half of 2015. Each of the six passed all 33 servicing metric tests for the first six months of the year, according to a December 2015 NMS monitor's report.
The problem with silos
The long nature of the servicing cycle--up to 30 years--can lead to mountains of data. That, combined with the scrutiny from regulators over longer periods of time, highlights the need for all systems to be interconnected. Data should flow freely between systems.
Yet even if the systems involving customer data and investor information talk to one another, people do not always communicate with each other. When individuals within servicing organizations fail to communicate necessary information to the right people, it can lead to siloed information, processes and decisions.
In fact, many servicers do work in an environment where processes and procedures are relegated to specific groups that may not have a full understanding of the role of other groups and how that affects their processes.
Different teams take responsibility for loans at different stages for different trigger reasons. Each team may not have a clear line of sight to the last team's function and process. All too often, this absence of insight and communication can lead to redundant or contradictory efforts by the various teams, with flawed results.
This lack of communication creates silos in servicing processes such as collections, bankruptcy, real estate-owned (REO) and loss mitigation. These problems are almost never contained to one group; they are the result of many variables that cannot be fully understood when there is little communication between groups.
A number of issues still exist relating to borrower communications--letters that don't conform to guidelines, poorly constructed scripts for phone interactions and lack of a single point of contact.
Staffing can be another issue. Regulators are looking for servicers to have adequate staffing with regard to quality control. Rules outlined by regulators need to be mapped to tests, and adequate--and experienced--staff must be on hand to manage testing.
While fixes within one group may solve a problem in that group's process, they can frequently trigger new problems in other areas of the platform and the company as a whole. For example, a fix within collections may end up causing a problem in loss mitigation.
What's needed is often an overhaul of a lender or servicer's approach to quality management. Following are some of the measures we've helped clients implement.
Quality framework is essential
The quality management framework is probably the most important piece of an effective overall quality program. A good framework implemented consistently across an organization makes execution more efficient, saving both time and costs. However, we've observed that very few organizations have a standardized methodology in place.
For example, many maintain multiple rules databases within the same company, with different departments interpreting the rules in different ways.
The quality framework should not only address regulatory requirements, but investor credit requirements as well. The framework should allow for measuring the quality of loan performance data, the effectiveness of credit policies and quality benchmarking.
An effective quality-management framework encompasses three primary principles: scoping, monitoring and verification. It follows six phases; rule management, risk assessment, test design, testing, reporting and remediation.
For now, let's focus on testing--an area where there is room for improvement.
Moving from reactive to proactive testing
Lenders and servicers design and implement quality tests to verify that servicing practices are in compliance with regulations and investor requirements. At the end of each reporting cycle, test results should be reported to both internal and external stakeholders, and any instances of non-compliance need to be corrected by the department in question.
As issues are identified, the root cause for errors needs to be determined and data should be analyzed so that all similar items can be examined. For example, if a particular lawyer, real estate agent, mortgage broker or appraiser shows a pattern of causing a loss or is involved with a questionable loan, all loans touched by that individual should be examined.
Today most companies have quality-management frameworks that meet the minimum need for investors and regulators. Yet we see significant limitations in the current approach to testing.
Rather than being performed early on, testing is still typically done later in the process and is more reactive than proactive. We see companies spread testing across departments or functions, from operations to oversight to auditing. Doing so adds costs and reduces the ability to cross-leverage between those functions. Today's testing processes are limited to the manual observation of defects, so errors are tracked, but real-time analysis is missing.
Ultimately, the testing systems in place today aren't efficient or cost-effective, resulting in defects that could have been prevented. That's because issues are identified after the fact, when losses have already occurred. Needless to say, remediation costs are expensive.
Testing can be streamlined by placing automated testing technology in the hands of the frontline operations (see sidebar, "Automation of Quality Tests"). By doing so, issues are identified and resolved much earlier in the process, potentially prior to a defect or problem occurring. Instead of reviewing individual transactions, control organizations should instead focus on reviewing the effectiveness of the automated controls.
Simply put, automating the testing of loans is a more effective and proactive approach to preventing defects. It's also more comprehensive than manually testing--the entire population of loans can be tested, rather than just a sample. Sampling risk is inherent whenever testing is performed on less than the full population.
By eliminating the majority of manual effort, an automated process ends up being cheaper. Defects can be readily analyzed, giving management access to better information so they can make better decisions.
Shared services in the mortgage operations structure
To achieve optimal value for an organization, there needs to be an effective use of resources. I recommend a quality shared services model, which consolidates quality activities that are traditionally performed by a number of control organizations throughout the business. A quality shared services model can help achieve a higher level of quality standards and improve decision support at a lower cost level. I see some larger mortgage lenders starting to adopt certain parts of the shared service model, but I believe the industry still has a long way to go.
Control organizations should share their technology infrastructure and common services where possible. All control organizations undergo a similar process from rule management to remediation. The technology requirements of each should not differ substantially. It is important that control organizations perform their functions independent of each other.
Verification activities, such as test design and testing, should have minor operational leverage. Scoping and monitoring activities, such as rule management and risk assessment, however, may be shared. The management of rules, as well as connections to policies and procedures, and assessments made at the scoping phase helps ensure that efforts are not being duplicated. It also helps ensure that the direction taken by all is consistent with overall enterprise objectives.
The results of verification activity by all groups should be consolidated and reported to management to enable better oversight, as well as insight into the entire operations as a whole. As findings by various groups could often be rooted to the same issues, remediation efforts assisted by change management and project teams should be managed centrally.
The shared services center should operate separately, with well-defined processes supporting multiple units and interacting with stakeholders of the organization, including line of business, control organizations, investors and regulators. This arrangement allows control organizations to focus on carrying out their core objectives. In addition, proper governance over the control organizations and shared services centers is important to ensure that the duties and responsibilities of each is well defined, do not overlap and are monitored to maximize the value of the model.
Taking a step back
Compared with many other industries, quality management is relatively new in the mortgage business. It hasn't been at the forefront in the mortgage sector in the way that quality control has been with, say, car manufacturers.
In many cases, there isn't a consistent approach to managing mortgage quality within the same organization or even the same department. Banks and other lenders have been so busy reacting to the myriad directives from regulators and investors that they've often ended up with disjointed and largely ineffective quality-management programs.
Yet more and more today, organizations are taking a step back to evaluate their approach to quality, and to implement more comprehensive and consistent quality programs. Progress is being made, and eventually the mortgage industry will catch up with some other industries when it comes to managing quality. And, of course, that's a very good thing for both lenders and consumers.
The IRG Model Is Here for the Long Term
The National Mortgage Settlement (NMS), reached between five of the largest mortgage loan servicers and 49 state attorneys general and the federal government, outlines 304 new servicing standards that the servicers are required to follow. It covers the servicing of loans, loan modifications and other forms of assistance to eligible homeowners.
Today servicers are recognizing that some requirements mandated by the NMS provide real value to their organizations. One of those mandates was the creation of an internal review group (IRG).
The NMS called for each servicer to establish its own IRG, which is charged with monitoring the servicer's compliance with the servicing standards. The agreement called for each servicer's IRG to be independent from the servicing line of business and to fall under either credit risk or an internal audit group.
The NMS outlines 33 enforcement metrics that are tested monthly and reported to the Office of Mortgage Settlement Oversight (OMSO) quarterly. Consistent errors on a metric by any of the servicers will cause them to incur fines that increase if not corrected.
Under the settlement, the servicers were required to provide more than $20 billion in consumer relief. Each servicer's IRG is charged with calculating and validating its progress toward meeting that institution's consumer-relief credits.
For the servicers, the challenge has been to create an IRG as a completely independent entity within their organization that would monitor the implementation of the NMS servicing standards over a three-year period.
To accomplish this task quickly, the servicers had to decide on the best approach to do the following:
* Determine staffing levels and staff requirements, based on their servicing portfolios;
* Identify the systems that needed to be accessed to perform the 33 metric tests, as well as the consumer-relief validation; and
* Decide on the technology that would be implemented to perform the required work.
Let's examine those three tasks and the options currently used by the servicers.
Determining staffing levels
Servicers first looked at the size of their mortgage servicing portfolios, which defined the quarterly sample populations under the terms of the NMS. The servicers then needed to assess whether they had the staff and expertise internally to perform the work.
The servicers took a few different routes when it came to staffing. However, it appears that a blended approach using a combination of in-house staff along with third-party vendors may be the best course in order to ramp up quickly. That's because few, if any, of the servicers had all of the staff and expertise internally to build an effective IRG.
Identifying the systems that needed to be accessed
The servicers needed to identify all Internal servicing systems that the IRG team would need to access in order to perform the tests and collect data. The tests and the data were used to develop evidence packages that are submitted to the secondary professional firm (SPF) to be audited and signed off on by the OMSO.
This was probably one of the more difficult tasks for the IRG to undertake. Complications arose in identifying data and systems when a servicer had multiple systems of record (SOR). Having a single SOR with ancillary support systems helped to reduce testing times for the IRG.
Defining the technology
Servicers needed to decide whether the IRG would invest in a third-party testing tool; leverage internal systems such as governance, risk, control (GRC); or use a spreadsheet approach.
My company observed all three approaches, as well as a combination of the three. All three have their advantages and disadvantages.
One of the more effective is the investment in a third-party vendor tool specifically designed for the IRG, which allows the servicer to customize the tool for the purpose of testing and evidence gathering. However, this can be costly in the short run if the servicer is not planning to continue with the IRG platform at the end of the three-year period.
After working with the NMS and servicers, it appears the development and implementation of an IRG provides real benefits to servicers. The IRGs offer long-term cost savings due to improved efficiencies. The IRG model also reduces servicing and default-management errors, ensuring compliance with increased regulation, which ultimately allows the servicers to better serve consumers.
The IRG model can be successfully replicated and can also be leveraged for other lines of business, within banks and non-banks. It's a viable model that servicers should maintain even after the terms of the settlement have been met.
Automation of Quality Tests
What do all these things have in common: fraud detection and security; monitoring of trading activities; and customer segmentation and targeted product offering?
The answer is they are all activities that financial services institutions have made significant improvements to in recent years due to advances in big data technology.
The mortgage industry overall has been slow to recognize the benefits of big data technology, and in particular how it can improve the way quality testing is conducted.
Today a significant portion of quality tests, although simple to execute, continue to be manual in nature. Testers spend a majority of their time extracting and reviewing information from scanned documents and systems of record. Not only is the manual approach laborious, it is also very costly.
Now, imagine if all documents requiring review are converted electronically from images of typed, handwritten or printed text into machine-encoded text and indexed on a data warehouse. Quality testers may perform a systematic bulk retrieval of records associated with the scanned documents.
Instead of manually extracting information from the systems of record, quality testers should systematically query records from the data warehouse. Although this sounds obvious, it is not uncommon for organizations to use the manual approach as historical data may be archived, limiting data availability.
Data-extraction techniques--combined with advancements in data analytics technologies such as machine learning, sentiment analytics and text analytics--have the potential to significantly reduce testing time frames and complete review tasks at a fraction of today's costs.
For example, the customer service/call-center environment is increasingly under scrutiny from regulators for unfair, deceptive, or abusive acts or practices (UDAAPs). There is no checklist or road map to follow to evaluate compliance. Only experienced staff are able to adequately discern compliance with such laws.
Big data technologies can be used to analyze unstructured data such as customer emails and call scripts for compliance with principal-based requirements. Algorithms are trained to recognize patterns within the data set, whether a communication with the customer passes or fails the requirement. The technology essentially leverages the knowledge of the experienced reviewer.
Jeffrey Hulett is a managing director in KPMG LLP's Mortgage and Consumer Lending Advisory Services practice in McLean, Virginia. He has more than 25 years of industry experience, and his background spans across mortgage, consumer and small-business lending products. He can be reached at firstname.lastname@example.org.