Printer Friendly

How do you measure up?

An annual study of servicers' costs showed the average direct cost of servicing dropped by 8 percent last year, on a per-loan basis. Although economies of scale showed up in the findings, the direct servicing costs per loan for midsized versus very large servicers showed little difference.

In 1992, the mortgage servicing industry focused heavily on benchmarks--and spent a great deal of time trying to achieve them. Despite many factors holding back their performance--most notably, a record-high prepayment rate--many mortgage servicers achieved new levels of performance.

If these servicing operations had merely set their sights on the benchmarks of 1991, they would not have performed up to par in 1992. The industry's performance last year reaffirms the aphorism that, "If you always do what you always did, you will always get what you always got." And almost always in servicing, that is simply not good enough. In mortgage servicing, staying in front of your competition today means setting your sights on tomorrow's benchmarks.

The 1992 KPMG Peat Marwick Mortgage Servicing Performance Study (MorServ) looked at the mortgage servicing industry a bit differently than in previous years. First, we split the group of participating servicers into two subgroups--the major market servicers, those with servicing portfolios of more than 100,000 loans, and the middle market servicers, those with servicing portfolios of 100,000 loans or fewer. In prior years, the study was only performed for large, major market servicers with more than 100,000 loans. The analysis that resulted from the addition of the middle market servicers was enlightening.

Next, we adjusted what we included as the components of direct cost to better reflect the way servicers measure their own operations. We included general and administration costs, data processing and record retention components in indirect costs. (All of the prior year results presented here have been adjusted for comparability.)

Finally, we increased the scope of the study, adding two new sections: one on technology-related information, the other on compensation. These sections are the result of demands for specific, validated information in these areas.

As part of the MorServ process (a proprietary and copyrighted process), we collected, validated and analyzed both quantitative and qualitative data on all areas of mortgage servicing, including customer service, investor services, risk management and miscellaneous functions. The scope of the study also includes indirect costs, non-operating costs (interest expense) and loan losses.

The validation process ensures that, to the extent possible, all results are presented on an "apples-to-apples" basis, and the inherent differences in the individual servicing organizations are noted and recorded. The goal is to produce comparable, reliable benchmarks covering cost, productivity and profitability. These data can then be used as a baseline that servicers can adapt to build their own benchmarks and goals against which to measure their performance.

The study specifics

The study covered approximately 5 million residential mortgage loans totaling $375 billion in principal balances. The servicers ranged in size from 25,000 loans to more than 1 million and were diverse in regard to geography, investor type, product type and level of automation. Approximately two-thirds of the participants are subsidiaries of financial institutions; the remainder is a mix of banks, savings and loans and stand-alone mortgage companies.

The data were analyzed along functional lines and stratified into 16 specific "departments," as well as by line-item detail, focusing on personnel (salaries, benefits, overtime, etc.) and other direct costs (outside services, occupancy and so forth).

The cost data that resulted from this year's study illustrated some of the changes occurring within the mortgage banking industry. The direct cost per loan decreased from $63 per loan in 1991 to $58 per loan in 1992, an 8 percent decrease.

Surprisingly, there was very little difference between the major and middle market servicers in regard to direct cost. The larger, major market servicers averaged $57.83 per loan, and the middle market servicers averaged $58.47 per loan. The middle market servicers had higher costs in the investor services area, where we would expect to see some economies of scale come into play, but they also experienced lower costs in risk management. Overall, however, the two lowest, direct-cost servicers were both major market servicers, whose portfolios had an average size during the year exceeding 300,000 loans.

When analyzing the study participants' total cost per loan (including direct cost, indirect cost, non-operating costs and loan losses), this cost figure was $195 per loan, a slight decrease from 1991. These results are more favorably weighted toward the middle market group, primarily because of high loan losses evident in the study findings for the major market servicers.

A number of the larger servicers had sizable loan losses as the result of recourse and/or owned portfolios, primarily located in California or the Northeast. Loan losses averaged $97 per loan for the major servicers, compared with only $32 per loan for the middle market servicers. In many cases, the middle market servicers are more regionally focused and did not have large percentages of New England and/or California product, allowing them to outperform the major market group in this area.

The only significant, consistent economies of scale between the two groups were realized in indirect costs, which include data processing, record retention and general and administrative expenses. General and administrative expenses include all servicing overhead costs, such as executive office, accounting/finance and human resources. Non-operating expenses, including prepayment interest expense, interest on escrows and interest on advances, were fairly consistent between the two groups of servicers. Compared with 1991, indirect costs decreased by approximately $2 per loan and non-operating costs decreased by almost $5 per loan, based on the major market servicer results.

The productivity angle

When analyzing any servicing operation, productivity and cost should be measured simultaneously. Productivity may be influenced by a number of factors, including outsourcing and automation, while cost is heavily influenced by the cost of living in a specific area.

We focused on various measures of productivity, including global and departmental productivity measures. By our definition, loans serviced per full-time equivalent (FTE) is a fully loaded productivity measure. We include direct and indirect FTEs, full time and part time, as well as including in our calculation such items as temporary help and overtime.

This year's average productivity findings did illustrate that the bigger, or major market servicers (636 loans per FTE) outperformed the middle market servicers (569 loans per FTE) in this key area. Interestingly, this year's middle market group of servicers was on par with the major market group in the 1991 study. Overall productivity appears to be increasing for all mortgage servicers.

When the calculation focuses on loans serviced per direct FTEs, the variance between the results of the two groups is more pronounced; the major market group averaged 779 loans serviced per direct FTE versus 675 for the middle market group. However, the middle market servicers in the study outperformed their major market peers in productivity in the indirect cost areas.

Obviously, there really is no such thing as an "average loan." This is why other productivity benchmarks were focused specifically on activity levels (i.e., adjusted for the variances of activity among the servicers). For example, we measured and analyzed incoming calls per customer service operator, tax payments per FTE and average delinquencies per FTE. In this way, an organization can have a standard to fairly evaluate the performance of the personnel in each area.

The compensation issue

As part of the compensation section, we used a cost-of-living index for each participant's servicing operation based on the ACCRA Cost of Living Index published by the American Chamber of Commerce Researcher's Association, Louisville. The index was used to adjust the compensation data that were collected for a specific job function classified as management, supervisor, senior-level staff and lower-level staff. This index also was used to help determine the effect of personnel costs, specifically that which differs based on the cost of living in the servicer's location.

On average, personnel costs accounted for 60 percent of direct costs. This included salaries, benefits, bonuses or incentives, overtime, temporary help and other personnel costs (relocation, auto allowances, etc.). The cost of living for the major market group averaged 1.05, with 1.00 representing an average cost of living in the United States. The middle market group of servicers had a slightly higher cost of living, averaging 1.07. The cost of living indices for the entire group of study participants ranged from 0.91 to 1.32.

Although those servicers in higher cost-of-living areas usually exhibited higher personnel costs per loan, they did not have higher overall costs per loan. The effect of the servicer's location on cost was evident but, on average, was not a significant driver of direct cost per loan. In the results for both the middle market servicers and their larger counterparts, the highest direct cost servicers had below-average cost-of-living indices.

Portfolio characteristics of note

Of the various portfolio characteristics that we examined, one of the consistent indicators of both direct and total cost in 1992 was delinquency rate. For the major market servicers, risk management made up 31 percent of direct costs and, including loan losses, 52 percent of total cost. This group's delinquency rate averaged 5.3 percent, a slight increase compared with the 1991 average of 5.10 percent.

In most cases, servicers with delinquency rates that were materially higher than the average also had higher risk management and direct costs. In many cases, those same servicers had higher loan losses because of loans sold with recourse or loans that had to be repurchased. As expected, those servicers with higher delinquency rates frequently experienced higher-than-average foreclosure rates.

Usually, an inverse relationship is noted between costs and productivity. For example, in risk management, participants with higher delinquency rates experienced higher costs with lower productivity. Productivity, as measured by total loans per collections department FTE, averaged approximately 5,500 on a combined basis for both groups of servicers.

As mentioned earlier, we also measured activity-based productivity benchmarks, such as average delinquencies per collections department FTE, as a productivity benchmark for collections. It was interesting that some operations that appeared very productive based on loans per collections department FTE were actually less productive than average based on delinquencies per collections FTE and vice versa. For this reason, activity-based benchmarks should be used in conjunction with more typical, portfolio-based benchmarks in order to determine achievable yet aggressive goals for a specific servicing operation.

The portfolio mix, mostly a matter of chance, has a great impact on servicing cost. The effect of such criteria as investor type (government, agency, private, parent or owned), percentage of fixed-rate versus adjustable-rate mortgages and year of origination should be considered when evaluating a servicing operation's cost efficiency and productivity, or when establishing benchmarks for future performance.

For example, in our examination, we noted that most participants with high percentages of GNMA servicing reported lower prepayment rates than their counterparts with greater percentages of conventional loans. Additionally, servicers with higher percentages of GNMA servicing experienced higher delinquency rates, which is consistent with prior year data.

Looking at all the participants' portfolios as a whole, 24 percent were GNMA portfolio loans, yet GNMA delinquencies accounted for 34 percent of all delinquent loans. Obviously, an appropriate benchmark for a servicer with 5 percent GNMA servicing may not be applicable to a servicer with 50 percent GNMA servicing.

Further, having large percentages of private investor loans generally corresponds to higher costs and lower productivity in the investor reporting area. This was evident for most servicers in both the major and middle market studies. However, the amount of standardization, the reporting requirements for particular private investors and the actual number of private investors can offset this general observation. Activity-based productivity measures in these areas, such as private investor loans per investor reporting department FTE, are often helpful benchmarks for evaluating actual productivity.

Another aspect of the portfolio mix that affected servicing performance--the impact of the fixed-versus-ARM mix--was evident in regard to both cost and productivity in various customer service areas. Costs in the special loans area, which includes ARMs, graduated payment mortgages (GPMs) and FHA 235s, were higher for servicers with larger percentages of ARMs. (ARMs usually make up the bulk of all special loans.) This is normally due to the additional tasks required when servicing ARM loans, including monitoring various rate indices, making ARM adjustments and performing ARM audits.

We would expect a greater number of customer service calls per loan for servicers with high concentrations of ARM loans. Although this might be the case, we could not isolate ARMs as the single criterion that caused the variance in the number of customer service calls.

Prepayments and payoff costs

The interest rate environment, specifically the prepayments that resulted from very low interest rates, affected all servicing operations in 1992. The annual prepayment rate in 1992 for the two groups averaged approximately 21 percent, compared with 12 percent in 1991, and 8 percent in 1990. However, the increase in prepayments had a more severe impact on the rate of churning and the decrease in portfolio size than it had on the cost of servicing in the payoff area. The payoff cost per loan increased only 28 percent, to $4.50, despite the 75 percent increase in the prepayment rate.

Although the average prepayment rates were virtually equal for the major and middle market servicers, the payoff cost was considerably higher for the smaller servicer group. This middle market group averaged $1.55 more per loan in payoff costs compared with the major market servicers. This may indicate that the larger servicers either had some excess capacity in this area prior to 1992 or that, because of cross-training and/or automation, larger servicers had to rely less on temporary staff, who are generally more expensive and less productive.

The shrinking of portfolios

The dramatic effect of the prepayment rate on average portfolio size provided a challenge to most servicing executives in 1992. For the major market servicers--many of whom are very aggressive in regard to portfolio growth through production and servicing acquisitions--the average change in portfolio size in 1992 was a decrease of almost 20,000 loans. This is a significant change compared with 1991, when the average portfolio increased by more than 8,000 loans.

Again, the lower number of loans in the average portfolio caused by the payoff rate, in addition to the extra work created by the volume of prepayments, may have worked to understate or offset the performance improvements that have occurred in the industry. When the prepayment rates return to a more normal level, the benchmarks of the industry may rise to indicate even higher levels of performance.

Profitability factors

When observing the profitability of a servicer, the portfolio mix obviously affects the average loan balance, which in turn affects the servicer's revenue and fee structure. In most cases, government loans, well-seasoned loans and loans from less populated geographic locations will carry lower principal balances. In the study of major market servicers, the net servicing fees averaged 33 basis points, or $254 per loan based on an average loan balance of $76,000. That compares with the middle market servicers' average of 30 basis points, or $212 per loan based on an average balance of $71,000.

Net servicing fees dominated the revenue area, accounting for 75 percent of total revenues. For both study groups, participants with higher average balances commanded more servicing fee dollars. Ancillary income, the largest component of which is late charges, was also affected by average loan balance, although in many cases other factors, including percentage of delinquent loans and percentage of late charges waived, also played a role. Ancillary income accounted for 13 percent of total revenue.

On average, the core components of revenue (net servicing fees, subservicing income and ancillary income) averaged $288 per loan, or 39 basis points. These results did not materially differ between the two servicer groups. Based on this figure, excluding any gain/loss on the sale of servicing rights, gain/loss on the sale of REO and interest income from the revenue calculation, and also excluding interest expense and loan losses from the expense calculation, the net profit margin averaged $196 per loan, or 26 basis points.

The results of this study provide a peer analysis of comparable cost, productivity and profitability data. In most cases, our goal is to identify the factors (investor type, delinquency rate, geographic location, degree of automation, etc.) that affect the results, but not to adjust the results for these items. All servicing operations are unique to a degree in their mix of characteristics. These are the characteristics that should be reviewed and analyzed before establishing any benchmarks or performance criteria.

We want to emphasize to both participants and non-participants that any peer statistics should be used as a baseline of information. Such numbers provide a starting point from which to determine appropriate benchmarks for a specific servicing operation for a specific point in time. These differences among servicers are critical when trying to determine if your operation is "up to par."

As is the case when developing any benchmarks, the process should be fluid to adjust for nuances and unexpected occurrences in the industry and economy. This way, as the mortgage servicing industry continues to change, a servicing operation will not be left pursuing goals that have since become obsolete. This ability to keep up with the industry's pacesetters, or to set the pace yourself by constantly improving performance, will determine success in mortgage servicing during the coming years.

Geoffrey A. Oliver is a partner at KPMG Peat Marwick. Regina J. Reed is a senior manager with the firm. Both are with KPMG Peat Marwick's National Mortgage Banking Practice, headquartered in Washington, D.C. The Mortgage Servicing Performance Study and Mortgage Production Performance Study are performed annually.
COPYRIGHT 1993 Mortgage Bankers Association of America
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 1993 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Title Annotation:costs of mortgage loan servicing
Author:Oliver, Geoffrey A.; Reed, Regina J.
Publication:Mortgage Banking
Article Type:Cover Story
Date:Jun 1, 1993
Words:2978
Previous Article:Wall Street's costly quest.
Next Article:The servicing watch list.
Topics:


Related Articles
Crossing the rubicon.
Reexamining 1994.
Protecting against employee dishonesty.
A new servicing cost study.
Integrating the e-Business Model.
Meridian Capital Group.
Refi Boom Strains Web Performance, Customer Service.
Bricks-and-Clicks Firms Gain Momentum.
Georgia and beyond: a predatory lending update. (Cover Report: Legislative/Regulatory).
Survey-happy.

Terms of use | Copyright © 2016 Farlex, Inc. | Feedback | For webmasters