REQUIRED READING: With today's increased scrutiny on the internal operations of lenders and servicers, it is only natural that the requirements for quality control programs have followed suit. Quality control (QC) programs monitor the risk associated with failures in the operational process. Investors and regulators closely analyze the QC process to ensure that loans being produced are of investment quality.
Unfortunately, many QC programs have not kept pace with the changes resulting from emerging regulations and loan evaluations in the mortgage industry.Â
Traditionally, the industry has followed the QC program requirements designated by Fannie Mae and Freddie Mac for conventional loans and Federal Housing Administration/Department of Veteran Affairs requirements for government loans. During the past year, those guidelines have undergone major revisions. For example, Fannie Mae's new guidelines place the responsibility for establishing quality standards on the lender rather than on the government-sponsored enterprises (GSEs) and investors.
Fannie Mae's new guidelines require lenders to develop and maintain a QC program that defines standards for loan quality and establish processes designed to achieve those standards throughout the entire origination book of business. The program must include the reporting of results of quality reviews to the lender's senior management, who should take actions addressing and remediating defects discovered in the lender's review process.
Sadly, lenders have been slow to respond. In large part, lenders remain committed to their existing programs and are not adapting their processes to meet the new requirements. Even worse, most lenders are not able to isolate the specific elements of their program that will cause an issue for investors and/or regulators. In other words, they don't know if their program has a problem.Â
In a proverbial nutshell, you know your QC program has a problem if your company experiences any or all of the following situations:
After reading your QC reports, you don't know if you have a quality problem or not. Most QC reports provide a list of the files reviewed and the findings for each individual file. This type of reporting fails to focus on the processes that created the issues in the first place.
It is critical for senior management to receive reports that clearly identify an overall percentage of the loans reviewed that failed to meet the company's defined quality standards. This is the first indicator that the established processes are not producing the level of quality that is expected.Â
One measure of quality may require that a loan file include all of the documents used to underwrite the loan. In this case, the QC report should list the number of loans that do not have all the necessary documents. The report should also show this number as a percentage of the total loans reviewed. If 100 files were reviewed and two were found to be lacking all of the loan documents, then the error rate is 2%.
When establishing the overall quality definition, senior management should also consider the costs associated with failing to meet their standards. The cost to produce a quality loan may be $1,000. Yet, the cost associated with a loan that does not meet the lender's standards could be double that amount when the costs of corrections or the need to sell the loan at a reduced price are taken into account. This knowledge can be used to determine the financial risk associated with specific error percentages and allow a more structured approach to managing quality.
The reports don't provide a confidence level. The results contained in the QC report are based on a sample of loans closed or purchased for the month. While there are various types of sampling programs available, the report should clearly identify how confident the QC department is that these results actually reflect the population of loans being reviewed.
Confidence in the result is a measure of certainty. For example, the confidence level required by the GSEs is 95%. In other words, the writer of the QC report is confident that if 100 samples were selected, 95 of those samples would have similar results.
However, lenders also need to allow for some variation in their findings, known as the precision rate. The more precise the results are, the less variation will be found. The agencies require that the precision rate be 2%, which means that the error rate may fluctuate up or down by 2%.
This information is critical to management if they are to make well-informed decisions on what action, if any, should be taken as a result of the QC findings. If the QC results show an unexpectedly high error rate but fail to define the confidence level of the results, management will have a difficult time determining if something needs to be changed or if this error rate was the result of a sample that randomly pulled in a high number of bad loans.
The QC reviews frequently try to address this problem by taking large samples of riskier loans. While a good idea conceptually, the failure of the program to select a sample based on a true stratification of the loans by risk attributes usually results in findings that are meaningless to management.
The reports do not provide meaningful comparisons using the company data, nor are they able to provide any type of comparative analysis for the industry as a whole. Some of the most meaningful information found in QC reports compares the results to previous months. Valid comparisons require that apples be compared to apples and loan populations be compared to similar loan populations.
However, if the sample is biased or the results are artificially based on pre-selected loan attributes that differ each month, any attempt at a comparison is invalid. Comparing the results of a random sample one month to a sample that is of a high percentage of a specific risk attribute the next month tells management nothing.
Most lenders have never been shy about stating that they produce ‘the highest quality loans,’ but this statement begs the question: How do they know? Currently there's no industry standard definition of a ‘quality loan’ that clearly outlines the attributes of a high quality loan. Setting clear standards would provide an effective means for better quality pricing.
Also, in order for standards to be meaningful, an independent database that is organized and fed by all lenders should be established. Once this is accomplished, management will have access to valuable QC reports that provide overall comparisons to the industry.Â
The results don't identify the processes or activities that created the problem. While today's QC programs differentiate findings based on the source of the files reviewed, all too often the reports don't focus on the processes themselves, and the information necessary to take action remains a mystery. A QC report is an operational review, not a loan audit. An operational review is focused on the processes and the activities within that process that combine to produce the expected result.
A good QC report should focus on processing or underwriting or closing, not brokers, loan officers or branches. If loan closings are failing to ensure that all underwriting conditions are acceptable, then this is a closing process problem, not a retail or correspondent lender problem.
There is no metric for measuring whether individual loans meet a company's expectation of quality. All QC departments should have a questionnaire or checklist to follow when reviewing loan files. The number of elements reviewed in a file can vary significantly. Some ask as many as 300 questions and others as few as 50. Yet, rarely do they ask the very basic question: Were operational processes followed to ensure the loan met its quality expectation?
The answer to that question is the critical measure of quality and leads to the analysis that provides guidance to management on how the process can or should be improved. Equally important is what drives the answer to this question. It is possible to have minor mistakes in a file and still have the expectation of quality met.
It is also possible to find loans that have only one or two errors, but those mistakes are significant enough to cause the loan to be defective, or fail to meet the definition of quality.
There is nothing in the findings to indicate whether the defects are random or systemic. While typical QC reports identify individual findings, and in some cases may actually group the findings, there is nothing that tells management if the issue is a random or systemic occurrence.
As is often the case, humans are prone to making errors. Holding training sessions to review processes where several random mistakes were made will not improve the overall quality of the process or the product. Instead, corrections should be focused on those issues that are found in a high percentage of the files.
The QC department should analyze why problems are occurring. Are the problems due to human error because the process is not understood? Or are problems occurring because the process itself is flawed, or there is a technology issue?
For example, one segment of a lender's processing department continually had a higher error rate than was acceptable. Despite a good deal of training and discussion to address the problem, the error rate did not improve. The QC staff then reviewed the process and activities required and found that current staffing levels were too low for the group to meet the expectations of the company. Once staffing was increased and some of the processes were automated, the overall number of errors dropped significantly.
The QC report should include this type of analysis along with recommendations for improvement. Otherwise, management may be wasting time and money on efforts that will not provide the desired return.Â
The review process is neither consistent nor objective. Humans conduct QC reviews in the mortgage industry. And no matter how hard we work on ensuring that each file is reviewed in the same manner or that each file is evaluated in its entirety, it is impossible to achieve the same level of consistency and objectivity as an automated program would. Lenders must ensure that, to the extent possible, QC activities and processes are automated and that the staff is trained and monitored for consistency.Â
Mid-level management has no interest in reading or responding to the reports. Since most QC reports delve into specific areas of origination or servicing, it is incumbent on managers in these areas to read the reports and respond to their conclusions. If mid-level managers are hesitant to spend time on this task, it is impossible for senior management to respond accordingly.
Responses that are aimed at addressing random issues or unspecified or unverified findings create a credibility problem for QC and senior management. Every report should reflect the confidence QC has in the results and clearly define the issues and the rationale behind the findings.Â
QC programs in the mortgage industry first emerged in 1985. Since then, lenders and servicers have, unfortunately, placed little effort or value in QC, and ineffective QC programs clearly contributed to the financial meltdown of 2008.Â
It is time for the industry to re-evaluate its take on QC. Effective QC programs are not just a regulatory burden but deliver real benefits to lenders and servicers. Properly designed QC control programs provide true value, as well as relief from excessive regulatory audits. The industry needs to break away from the problem-riddled methods of the past and demand that QC programs meet the same level of quality found in other sectors of a company.
Becky Walzak is president of Looking Glass Group LLC, an Indianapolis-based consulting firm focused on business transformation initiatives in the financial services industry. Walzak is also the president of rjbWalzak Consulting, based in Boca Raton, Fla. She can be reached at (561) 459-7070.