Servicing profits and performance.

Servicers continued to focus on holding down costs in 1996, a new study shows. Direct servicing costs edged up only slightly to $48 per loan compared with $47 in 1995. A new accounting standard clouds measurements of profitability, yet even excluding its impact, servicing per-loan profits of $233 failed to match the peak profit year of 1994.

If you are trying to evaluate mortgage servicing performance, especially in terms of profitability, the rules have changed. While reporting is not yet exactly the same among companies, the items that should be measured and how to calculate them are more consistent and complete among companies than they have been in the past. With the implementation of the Financial Accounting Standard Board's (FASB) Statement of Financial Accounting Standard No. 122, "Accounting for Mortgage Servicing Rights," companies evaluating their mortgage loan production, secondary marketing and servicing are changing the way they look at these components and their overall results.

The 1995 pretax return on assets (after amortization of mortgage servicing rights) of 8.4 percent decreased slightly to 8.1 percent for some of the best performers in 1996, based on the latest KPMG Peat Marwick LLP servicing performance study [ILLUSTRATION FOR FIGURE 1 OMITTED]. These returns are not in the range of 15 to 25 percent that executives have observed in the past, due to narrowing margins and a smaller asset base caused by the fact that previous accounting disallowed capitalization of originated mortgage servicing assets. As a result, KPMG in its 1996 annual mortgage servicing performance study, MorServ, addressed both the profit returns and operational performance of mortgage servicers to assess the opportunity for enhancing profits and returns.

Aside from industry consolidation, the relatively flat interest rates have caused little change in the participants' mortgage servicing operations. This was evidenced by the relative consistency year-to-year in the average direct cost per loan of $47 in 1995 to $48 in 1996 and average direct productivity of 1,044 in 1995 to 1,106 in 1996. However, the reported profitability trends for financial statement reporting have declined due to the impact of FAS 122. As shown in Figure 1, if you exclude the impact of FAS 122, profitability slightly decreased from 1995 to 1996 for the more profitable companies and declined by 2 percent for the remaining participant group.

1996 MorServ

This was the 11th year of KPMG's MorServ study. MorServ provides performance data that allows mortgage finance companies to compare themselves on an apples-to-apples basis with other mortgage servicers. To achieve comparability, the participants' data was thoroughly validated by KPMG to help ensure the data was presented consistently and accurately despite the differences that exist among the servicers' internal reporting measurements.

In addition to the financial performance analysis, MorServ focused on single-family mortgage loan servicing operational quantitative and qualitative data. Factors such as portfolio characteristics, technology use, management structure and compensation structure were compared against the profitability of the servicers to identify best practices within the industry.

Although no two portfolios are exactly alike, differing in such areas as average loan size, portfolio size, investor mix and geographic locations of the underlying mortgages, it is important to understand how these differences affect the profitability, cost and productivity of each organization. From the study results, KPMG concluded that companies with higher profitability were profitable because of the characteristics of their portfolios as opposed to the number of loans in their portfolios.

Many times a mortgage company identifies its competition based on its rival's portfolio size while ignoring many of its other competitors. Also, many servicers do not consider companies that perform servicing activities differently as competition, although they compete for loan servicing in some of the same markets. The traditional definitions of competition did not include profitability, which is the primary driver determining survivors in mortgage servicing today and in the future.

In producing this year's MorServ study, KPMG studied the data for the 12 months ending Dec. 31, 1996. This year's MorServ study included data for mortgage servicing operations with a combined servicing volume of more than $400 billion. The portfolios surveyed ranged from 45,000 loans to more than 1 million loans. Similar to 1995, portfolio composition averaged 50 percent conventional agency, 21 percent government, 20 percent parent/own and 9 percent private investors.

The ownership structure of the participant group was 55 percent bank-owned, 27 percent independent and 18 percent thrift/savings and loan. While the mortgage servicers in MorServ do not represent the average for the industry as a whole, they do represent a broad cross-section of the top 100 servicers.

Participants responded to an in-depth questionnaire that focused on profitability, cost and productivity. The survey included questions regarding activity levels, operational processes and statistics, technology and portfolio composition. As a result of KPMG's approach, some of the better mortgage servicing practices were determined. KPMG's analyses encompassed each of the critical functional areas identified in MorServ:

* customer service (customer inquiry, taxes, insurance, escrow analyses, payoffs and assumptions)

* investor services (investor accounting/reporting and cashiering/cash management)

* default management (collections, workouts, bankruptcies, foreclosures and real estate-owned)

* new loan set-up (electronically transferring or manually inputting borrower information into the servicing system)

* servicing acquisition (due-diligence activities associated with acquiring loans via bulk servicing purchases)

* indirect functions (general and administrative, data processing and record retention activities).

MorServ historical trends

While mortgage banking is a cyclical business and the participants in MorServ change annually, KPMG continued to observe the trends in the industry during the 11 years of the study for profitability, cost and productivity.

Profitability from 1991 to 1996 improved, excluding the impact of FAS 122. In 1991 per-loan profit was $218, in 1996 profits per loan were $233. However, the peak profitability reached in 1994 at $329 per loan was driven by bulk servicing sale gains of $66 per loan, float income of $64 per loan, and net servicing fees averaging 44 basis points. The decline since 1994 is attributable to lower servicing fees earned, not lower average loan balances and lower ancillary income [ILLUSTRATION FOR FIGURE 2 OMITTED].

Over the years, direct costs have generally decreased while productivity has increased. The exception to this trend occurred during the refinance era of a few years ago, when heavy refinancing promoted an increase in costs because of high payoff rates and high servicing-released sales. The 1991 average direct cost of $81 was much higher than the $48 average direct cost in 1996. This is because of a combination of factors such as enhanced technology, economies of scale and outsourcing [ILLUSTRATION FOR FIGURE 3 OMITTED].

Productivity also has improved. In 1992, the number of loans per direct full-time equivalent employee was 700 compared with 1,106 in 1996 for the same reasons identified for direct costs [ILLUSTRATION FOR FIGURE 4 OMITTED].

Direct costs continue to be dominated by customer service, which accounts for 48 percent of the direct costs for the most profitable companies, 42 percent for the remaining participant group, and 43 percent for the average database [ILLUSTRATION FOR FIGURE 5 OMITTED].

Economies of scale provided very little benefit in the customer service department where 40 percent of the department's time was spent on customer inquiry activities. While KPMG's analyses identified that the more profitable companies spend proportionately more than other companies in customer service, the benefit of this expense does not appear to translate into customer retention. Some companies claim retention rates of 25 to 35 percent, yet most companies retain less than 15 percent of their customers. There was no evidence that higher customer service costs translate to customer retention.

The primary question raised through KPMG's analysis was whether the cost of customer service was a good investment or merely a proactive measure to reduce possible future servicing problems. Profitability should be a consideration in management's decision to increase or decrease customer service costs. Factors to be examined in making this decision focus on fee income opportunities, customer retention and reducing the risk of operational errors.

Characteristics of the profitable servicers

The companies with the highest profits on a per-loan basis (excluding the impact of mortgage servicing right [MSR] amortization and gains on bulk sales of servicing) were evaluated to determine the characteristics driving profitability. Those companies classified as most profitable represented a combination of bank-owned and independently owned mortgage companies with average loan portfolios of $60 billion. Their portfolio composition by investor type was 47 percent Fannie Mae and Freddie Mac, 27 percent Ginnie Mae, 15 percent private and 11 percent parent/own.

Similar to the 1995 results, KPMG found that in 1996 companies that focus on managing their revenues and portfolio characteristics as aggressively as their direct costs and productivity realized higher mortgage servicing profitability than companies whose efforts were primarily focused on direct costs and productivity. The difference between the better performer's direct costs of $40 per loan and the average remaining participant group's direct costs of $52 per loan was significant. But that difference was not as large as the revenue difference of $72 per loan for these two categories of performers. The better performers saw revenue of $391 per loan as compared with $319 per loan for the average remaining participant group.

The drivers of servicing revenue in 1996 were float income and net servicing fees, which was consistent with 1995. While 1996 float income of $51 per loan for the more profitable companies was less than their 1995 float income of $73 per loan, it was still higher than the $44 per-loan float income for average remaining participant group in 1996.

Companies focusing on a larger number of loans with escrows were more profitable, as evidenced by 75 percent escrows on average for the more profitable companies compared with the 70 percent escrows for the remainder of the participants. The 1996 best-performer average loan balance per loan serviced of $86,000 did not have as large an impact on profitability in 1996 as it did in 1995 when the average loan balance per loan serviced was $103,000. But the higher percentage of government loans in 1996 (27 percent compared with 8 percent in 1995) did affect the escrow percentage, thus affecting float income. Therefore, two approaches to managing the float income, aside from sophisticated investing techniques, were evident:

* Loans with higher average loan balances resulting in lower escrow percentage but higher aggregate escrow balances.

* A higher percentage of government loans that carry lower average loan balances but a higher percentage of loans requiring escrows.

The other significant profit driver in 1996 was net servicing fee income, which includes guarantee fees. The most profitable companies realized net service fee income of 35 basis points, which was significantly higher than the remaining participant group of 25 basis points. As stated earlier, when comparing the investor mix of the portfolios, the higher percentage of government loans heavily contributed to profits with no impact on the cost of servicing for the most profitable companies. Average loan balance at $86,000 each was comparable between the most profitable companies and the remaining participant group. This translated to per-loan net servicing income of $299 for the most profitable companies and $214 for the remaining participant group.

Default management costs tended to be more costly, by approximately $5 per loan, for the average company than for the most profitable companies. This difference was not due to the average delinquency rate since the most profitable companies tended to have a higher delinquency rate than the average company. Rather, the most profitable companies were more automated and aggressive in their collections activities, which contributed to the reduced cost of default management.

Which products are more profitable to service?

To confirm the benefit of servicing government loans, KPMG compared the portfolio mix of companies in the study with their net profits (excluding amortization of MSRs and gains on sale of servicing). We found that companies whose portfolios contained more than 50 percent of government loans realized net profits of 46 basis points. Companies with portfolios that contained more than 50 percent conventional agency loans realized net profits of 20 basis points. Companies with portfolios split evenly between government and conventional agency loans realized net profits of 33 basis points [ILLUSTRATION FOR FIGURE 6 OMITTED]

Many in the industry believe the cost of servicing government loans is higher than the cost of servicing conventional loans. However, based on comparing companies whose portfolios contained more than 50 percent government loans, KPMG found that the total cost of servicing, including costs such as loan losses and interest on advances, was $121 per loan. Companies focused on conventional agency loans had total costs of $155 per loan. Companies whose portfolios were evenly distributed among government and conventional agency loans incurred total servicing costs of $106 per loan.

KPMG believes the cost relationship referred to previously is not driven solely by companies with large portfolios. The average number of loans in the servicing portfolio for companies with a government concentration was slightly more than 300,000 (thus approximately 150,000 government loans were serviced) with the smallest company servicing only a total of 45,000 loans. The conventional agency-concentrated servicers averaged a total servicing portfolio size of 275,000 loans with the smallest company servicing 85,000 loans. The companies with even distribution between a government and conventional agency concentration averaged 500,000 loans with the smallest company servicing 110,000 total loans. This fact further confirms that for companies classified as a "top-100 servicer" the portfolio characteristics drive the profitability more than the number of loans in portfolio.

MSR Impact on profitability

Clearly, the implementation of FAS 122 affected the profitability of mortgage servicing operations because it requires amortization of an asset that historically has been unrecorded. On average, in 1996, the amortization and impairment adjustment tended to reduce profits by $137 per loan based on the amounts companies actually recorded in their financial statements.

Because of the differences surrounding the amounts companies capitalize for MSRs and the amortization and impairment techniques used, KPMG performed a calculation to compare on a consistent basis the impact on the return on assets. This is depicted in Figure 1. Assuming consistent MSR asset capitalization and consistent amortization expense, this graph identified a reduction of profits of $140 per loan, or 12 percent in the return on assets due solely to FAS 122. The difference between the KPMG calculations and the actual amounts recorded represent a reduction in returns of approximately 6 percent on average, which was weighted more heavily toward companies classified as average (not in the most-profitable classification).

Interestingly, regardless of the concern over recording an asset considered relatively volatile by many in the industry, KPMG found that 20 percent of the participating companies in MorServ do not hedge their MSR asset. Also, of the 80 percent that hedge servicing, only 60 percent received hedge accounting treatment, rendering the usefulness of the hedge somewhat questionable from an accounting and reporting perspective. Finally, slightly less than 60 percent of the total MSR assets were hedged by those companies that said they were hedging MSRs. The typical hedge instruments used were floors, swaps and bonds, and usually one type of instrument rather than a combination of instruments.

On average, in 1996, the cost of hedging, including the gains/losses, was only $1 per loan serviced. However, the range of hedging costs on a per-loan basis was a gain of $30 vs. a loss of $19. The most profitable companies experienced hedging losses of $4 per loan compared with break-even (i.e., no cost) for the remaining participant group that hedged. Clearly, in 1996, the cost of hedging was not significant to the overall profitability of the operations for either the most profitable or the average companies.

Management issues going forward

The industry seemed to continue its focus on addressing rising interest rates and the impact on the mortgage production business more than the mortgage servicing business from an operational perspective. The activities in mortgage servicing seemed to be more strategic than operational, which, given the declining impact direct costs have on profitability, seems appropriate.

KPMG is most often asked about strategies relating to possible servicing portfolio acquisitions and sales, identifying niches for medium-sized mortgage servicers, subservicing activities and managing the impact of MSRs.

Unfortunately, overhead costs continue to be ignored by servicing and other senior managers because of their perceived lack of control over costs or acknowledgment of their impact on profitability. General and administrative costs represent about 18 percent of total costs to the average mortgage servicer. These costs are one of the few types of costs that are affected by economies of scale. While the servicing manager typically cannot affect this category directly, it certainly should not be ignored by senior management.

While not as significant, another indirect cost category that continues to be discussed by senior management in a strategic setting is data processing. Data processing accounts for 11 percent of total costs. KPMG's study showed that only the larger servicers are taking advantage of newer technology. Companies with larger servicing portfolios have an advantage over smaller companies in cost-effectively implementing newer technologies and workstations to help manage workflow and decision making. However, the newer servicing technology available today may reduce a small percentage of the total cost to service with no ability to increase revenues. The new technologies currently implemented by the larger servicers focus on managing defaults by moving a loan more quickly from nonperforming to performing or foreclosure to minimize losses.

Other technologies, such as imaging, require appropriate training and a controlled implementation plan to ensure the benefits of this technology are realized in a timely fashion. Too many times the process is not changed and thus the technology that's implemented causes a net addition to costs rather than a savings. Management then blames technology for the increase in costs rather than blaming incorrectly administered implementation plans.

While most economists predict rising interest rates during the remainder of 1997, senior management is likely to continue to focus on replacing run-off in the servicing portfolio. KPMG believes it is equally important for mortgage companies to focus on the profit drivers of their servicing portfolio. Management should identify the types of products a company can originate and service more profitably, and at the highest return, in order to make the proper strategic decisions for the company as a whole.

While our analysis showed that government loan servicing can be profitable, this is not the case for all companies. KPMG study results have found that companies that focus on servicing specific product types, such as conventional, government, adjustable-rate and fixed-rate loans, have typically been more profitable than those companies that believe volume is the sole profitability driver. To set appropriate strategies for the mortgage company and record the returns the industry expects, improved reporting to senior management for both production and servicing operations is necessary.

Geoffrey Oliver is the partner in charge of KPMG Peat Marwick LLP's Mortgage and Structured Finance Group, headquartered in Washington, D.C. Laura McDonald is a senior manager with the Mortgage and Structured Finance Group.