Printer Friendly

The MBA Cost Study: measuring performance.

THE MBA COST STUDY Measuring Performance

OVERALL, 1990 WAS ANOTHER YEAR of improvement in the profitability of mortgage servicing according to a survey conducted by the Mortgage Bankers Association of America (MBA). The MBA Cost Study, an annual survey, is based on detailed income statement data from mortgage banking institutions. The 1990 study was developed from information provided by 85 companies. The firms surveyed included companies of all sizes, ownership structures and geographic locations. The average dollar volume of single-family loans serviced by participants in the survey was $2.5 billion and the average number of single-family loans serviced was 42,183.

For the 1990 report, the individual company data were obtained from the Mortgage Bankers' Financial Reporting Form. Quarterly submissions of this form are now required from all mortgage companies that are approved GNMA, Fannie Mae and/or Freddie Mac seller/servicers. The form requests detailed allocations of income and expenses for mortgage companies by functional area (i.e., loan production, warehousing, marketing and servicing). The Cost Study provides analyses and detailed data on the revenues and costs associated with each of these four functional areas. This article examines trends in income and expenses related to the servicing operation.

Servicing remains profitable

Net servicing income, after reaching a low point in 1987, increased in each of the next three years. Moreover, in 1990, servicing was the most profitable aspect of the business at most mortgage companies. For the average firm in the survey, the profit margin for servicing was 29 percent in 1990, nearly 3 percentage points higher than the previous year and 10 percentage points above the 1987 trough (see Chart 1). In per-loan terms, recent developments were similar. Net income per loan serviced was quite high in the mid-1980s, then it took a nose dive, reaching a low of $37 per loan in 1987. Since then it has steadily improved. In 1990, net income per loan serviced was $69, 13 percent higher than the $61-per-loan figure for 1989 and quite similar to levels maintained in the early 1980s.

The key to the recent increase in profits continued to be the emphasis servicers are placing on cost containment. After peaking in 1987 and 1988, total servicing costs actually declined in both 1989 and 1990. Servicing expenses as a percent of the dollar volume serviced were only 28 basis points in 1990, compared to 31 basis points in 1991 and 35 basis points in 1987 and 1988 (see Chart 2). Moveover, total servicing costs were only $169 per loan in 1990, 1 percent less than the previous year.

Direct expenses, the largest component of costs, were about unchanged in per-loan terms in 1990, but fell from 16 basis points in 1989 to 15 basis points in 1990, when expressed as a percent of the dollar volume serviced. This decline in direct expenses reflected another increase in servicing productivity. The number of loans serviced per servicing employee increased by 13 percent to 784 in 1990, from 696 the prior year. Another major cost component, provision for loan losses, also declined in 1990. With delinquency and foreclosure problems improving on average, expenses allocated to provision for loan losses declined from 3 basis points in 1989 to 2 basis points in 1990. The third principal cost category, the amortized cost of purchased servicing (ACPS), was also lower in 1990. As a percent of the dollar volume serviced, the ACPS declined to 11 basis points in 1990 from 12 basis points in 1989.

Income falters

Partially offsetting these positive developments with respect to costs was a decline in servicing income. Gross servicing income rose 3 percent in per loan terms to $239 in 1990. However, this increase was due solely to a 10 percent rise in the average loan size serviced, from $54,786 in 1989 to $60,208 in 1990. Servicing income as a percent of the dollar volume serviced declined from 42 1/2 basis points in 1989 to 40 basis points in 1990 (see Chart 3). The decline in 1990 was consistent with a pattern of declines that began in 1984, and pulled servicing income as a percent of the dollar volume serviced down to its lowest level in more than a decade. It is likely that this steady decline reflects increased competition that is driving servicing fees down. In 1990, servicing fees were only 34 basis points, 2 basis points less than 1989, and 5 basis points lower than in 1988. Other loan administration income was also down in 1990. Other loan administration income was $33 per loan in 1990, 9 percent less than in 1989, and as a percent of servicing volume, it fell to 5 basis points in 1990 from 7 basis points in 1989. (The two components of income do not sum to the total due to rounding.)

In addition to the traditional sources of income from servicing fees, the selling of servicing rights has become a standard part of many servicing operations. Mortgage companies sell servicing for a variety of reasons, including the generation of cash to fund losses from problem loans, paying off long-term debt and restructuring their servicing portfolios. Estimates indicate that servicing transfers have risen steadily since 1985, when they were approximately $80 billion. In 1990, servicing transfers were about $190 billion compared to $155 billion in 1989. Consequently, proceeds from sales of servicing can have an important influence on income. As a percent of the dollar volume serviced, gains from sales of servicing for the average firm in the survey fell to 6 basis points in 1990 from 8 basis points in 1989. Including gains from sales of servicing, total servicing income increased 14 percent to $273 per loan. As a percent of the dollar volume serviced, it rose from 40 basis points to 45 basis points.


Servicing expenses can be broken down into two broad areas: direct expenses and nondirect expenses. Chart 4 shows that during the past decade, there have been some marked shifts in the proportion of total expenses devoted to each of these two areas. Direct expenses, which include costs for personnel, electronic data processing (EDP) and occupancy, among other things, fell from 80 percent of total expenses in 1980 to 53 percent in 1990. In contrast, nondirect expenses, which include provision for loan losses and the ACPS, rose from about 20 percent in 1980 to 47 percent in 1990.

Charts 5 and 6 show the percentage of total expenses devoted to some of the key cost components within direct and nondirect expenses in 1980 and 1990, respectively. The big changes occurred in personnel, provision for loan losses and the ACPS. Personnel costs were 45 percent of total expenses in 1980 but declined fairly steadily since. By 1990, personnel costs accounted for only 19 percent of costs. On the other hand, both provision for loan losses and the ACPS increased as a percent of total expenses. This was not surprising given that during the past decade there was a dramatic increase in delinquency and foreclosure problems, and servicing transfers became more and more common. Provision for loan losses rose from 1 percent of total expenses in 1980 to 17 percent in 1988. But, in the last two years, this cost declined and in 1990, was only 7 percent of the total. In 1980, the ACPS represented 19 percent of total expenses. By 1990, it accounted for 40 percent.

As noted, total servicing expenses expressed as a percent of the dollar volume serviced were 28 basis points in 1990, down 3 basis points from 1989. Both direct and nondirect expenses were lower, and within nondirect expenses, both provision for loan losses and the ACPS declined.

The servicing expense ratio - direct expenses as a percent of the dollar volume serviced - fell from 1988 to 1990 (see Chart 7). The servicing expense ratio peaked in 1987 at 21 basis points; by 1990, it had fallen to 15 basis points. In per-loan terms, direct expenses were $89 in 1990, about the same as the year before and only slightly higher than in 1988. The decline in direct expenses reflected steady increases in servicing productivity (see Chart 8). Productivity rose for four consecutive years from 1986 to 1990, and the 1990 level was the highest recorded in the history of the Cost Study. In 1990, the number of loans serviced per servicing employee increased another 13 percent to 784 loans.

Within direct expenses, several key cost categories declined in 1990, including personnel, occupancy and EDP. "Other" direct expenses constituted the only cost category to increase in 1990.

Personnel costs declined for the third consecutive year in 1990. As a percent of the dollar volume serviced, personnel costs fell to 5 basis points from 6 basis points, and in per-loan terms, declined 9 percent to $32.

Occupancy expenses were $4 per loan in 1990, 15 percent less than the previous year. Occupancy expenses per servicing employee declined to $3,441 in 1990 from $3,578 in 1989.

EDP costs also fell markedly in 1990. Per-loan-serviced EDP costs were $6 in 1990, 21 percent less than the previous year. This was the third consecutive yearly decline in EDP expenses. However, these drops did not necessarily mean the firms were cutting back on their investment in computers. On the contrary, it could have reflected the increased use of even more cost-effective technologies, such as optical disks for data management, power dialing and computer voice response systems, which improved productivity so much that per-loan costs fell. (It should also be noted that part of the decline in EDP costs may have been due to the fact that the average firm in the 1990 sample was somewhat smaller than in the 1989 sample, and smaller firms, on average, tend to spend less on EDP than larger firms.)

Other direct expenses, the only cost area within direct expenses to rise, were $47 per loan in 1990, 15 percent higher than in 1989. Almost all of the increase was in the catch-all, "all other direct expenses" category, which increased 33 percent to $36 per loan. This category includes assorted expenses for telephones, office supplies, postage, advertising, etc. Corporate interest losses on mortgage-backed security (MBS) pools remained at about $4 per loan and non-EDP depreciation was nearly unchanged at $1 per loan. Expenses to service foreclosures and REO declined to $5 per loan in 1990 from $8 per loan in 1989.

The cost of problem loans

After being almost non-existent in the early 1980s, problem loan expenses rose dramatically in the middle of the decade. But, for a second year in a row, problem loan expenses improved in 1990 (see Chart 9). Expenses related to problem loans, which include provision for loans losses, expenses to service foreclosures and REO, and losses on sales of REO, were only $17 per loan in 1990, 39 percent lower than the $27 figure recorded in the prior year. Moreover, as a percent of the dollar volume serviced, problem loan expenses fell to 3 basis points in 1990 from 5 basis points in 1989.

All three components registered declines in 1990. Provision for loan losses fell to $12 per loan in 1990 from $17 per loan the previous year. Expenses to service foreclosures and REO declined 43 percent in 1990 to $5 per loan, and losses on sales of REO were less than $1 per loan compared to $2 per loan in 1989.

This decline in problem loan expenses was mainly a reflection of the recent improvement in the delinquency and foreclosure situation. Data from MBA's National Delinquency Survey show that the percentage of loans 60 days or more past due fell from an average of 1.52 percent in 1989 to 1.45 percent in 1990. Also, the percentage of loans in the foreclosure process at the end of 1990 was 0.94 percent, 4 basis points lower than at the end of 1989.

More specifically, VA foreclosures, which can be very costly, declined 7 percent in 1990 to 38,039. Furthermore, the VA's use of the no-bid option fell as well. In 1990, the VA used the no-bid option in only 11 percent of its foreclosure decisions compared to 16 percent in 1989. With total VA foreclosures declining and no-bids being used less frequently, no-bids fell 52 percent from 6,641 in 1989 to 4,249 in 1990. Even including buydowns, the VA foreclosure situation clearly improved. The combination of no-bids and buydowns was 25 percent lower in 1990 than it was in 1989.

Given the increased importance of servicing transfers, it was not surprising that the ACPS rose throughout the 1980s, with the biggest jumps occurring during the second half of the decade (see Chart 10). In 1980, the ACPS was $10 per loan. By 1985, it had risen to $24 per loan, and in 1990, it was $69 per loan, 4 percent higher than the year before. Despite the rise in per-loan terms from 1989 to 1990, the ACPS fell slightly as a percent of the dollar volume serviced. In 1990, the ACPS was 11 basis points compared to 12 basis points in 1989. Finally, as was true in previous years, purchased servicing tended to be a more common growth strategy for large servicers. It seems that the larger the servicing portfolio, the higher the share that purchased servicing represented of the total portfolio. This was in contrast to the sale of servicing rights, which tended to be higher at the smaller firms. Putting this together, it appears that small firms are, on average, net sellers of servicing rights, while large firms are generally net purchasers.

Profitability by size and type

The Cost Study also examined the issue of economies of scale and once again found that firms with large servicing portfolios tended to have higher levels of productivity and lower direct costs. At firms with servicing portfolios of less than 5,000 loans, the number of loans serviced per servicing employee was only 370, while at firms with servicing portfolios greater than 70,000 loans, the number of loans serviced per servicing employee was 884.

As expected, these differences in productivity carried through to costs. Direct expenses declined from $158 per loan for the smallest servicers to $82 per loan for the largest servicers. Within direct expenses, these cost trends held true for most of the key costs components, including personnel, occupancy and EDP.

In 1990, as always, the profitability of the servicing operation varied by type of business structure. But, as has been the case for the past several years, the servicing portfolios of privately held mortgage companies were more profitable than those at firms with other organizational structures. Net income per loan serviced at privately held companies was $111 in 1990, 61 percent higher than the $69 figure for the average firm. Gross income was higher than at the average firm and direct expenses were lower. In addition, these firms were not active purchasers of servicing rights, so the ACPS was low.

Net income at bank-holding-company-controlled mortgage subsidiaries, at $75 per loan, was also higher than at the average firm. The key for this group was lower-than-average expenses.

Net income at thrift institution subsidiaries and non-bank, non-thrift subsidiaries were both around $50 per significantly lower than at the average firm. In both cases, the differences were on the cost side. At thrift institution subsidiaries, both direct expenses and provision for loan losses were extremely high. At non-bank, non-thrift subsidiaries, direct expenses were higher than at the average firm, as was the ACPS.

The study also looked at companies that provided data in both 1989 and 1990, focusing on the 10 firms with the highest overall profit margin. In 1990, these 10 firms clearly distinguished themselves in servicing. At these companies, total servicing income was 8 percent higher, while direct expenses were 24 percent lower than at the average firm (see Chart 11). Servicing income was $257 per loan at the high-profit firms compared to $239 per loan at the average firm. The average loan size serviced was quite similar for both groups, but servicing fees were much higher for the high-profit firms. At these 10 firms, servicing fees were 38 basis points compared to 34 basis points at the average firm.

On the cost side, direct expenses were only $68 per loan compared to $89 per loan at the average firm. Higher productivity accounted for much of the difference. For the high-profit firms, the number of loans serviced per servicing employee was 983, 25 percent higher than the average firm. Somewhat surprising, however, problem loan expenses were higher for these 10 firms. Provision for loan losses, the largest problem loan expense category, was nearly $15 per loan compared with $12 per loan at the average firm.

Finally, these 10 firms were active buyers of servicing rights. Consequently, the ACPS was quite high for this group. At the high-profit firms, the ACPS was $85 per loan, while at the average firm it was $69 per loan.

But, even taking the cost of purchased servicing into account, net servicing income for these 10 firms was $89 per loan, 28 percent higher than at the average firm.


The prospects for mortgage servicing in 1991 and beyond appear promising. Total servicing volume is likely to continue rising, as is the mortgage banking industry's share of the market. In 1991, the share of the residential servicing market held by mortgage bankers was estimated to be about 38 percent, 1 percentage point higher than in 1990. By the middle of the decade, the mortgage bankers' share could be as high as 42 percent.

In terms of the everyday operations of loan administration departments, the focus will continue to be on cost containment, and based on data from MBA's quarterly Ratio Analysis Survey, this emphasis is paying off. Through the third quarter of 1991, servicing productivity has continued to rise and direct costs have fallen.

Unfortunately, however, the delinquency and foreclosure situation appears to have deteriorated in 1991. Also, with servicing transfers now commonplace, the ACPS is likely to remain high.

On the income side, servicers are not likely to see any major changes in 1991. If the current trend continues, servicing fees in per-loan terms are likely to remain relatively stable. Also, the sale of servicing rights will continue to be an important source of revenue generation.

On the whole, even taking into account the potential rise in problem loan expenses, 1991 appears to be another year of improvement.

PHOTO : CHART 1 Profit Margin - Servicing Operation

PHOTO : CHART 2 Total Servicing Expenses

PHOTO : CHART 3 Servicing Income as a Percent of Servicing Volume

PHOTO : CHART 4 Shifts in Servicing Expenses

PHOTO : CHART 5 1990 Servicing Expenses

PHOTO : CHART 6 1980 Servicing Expenses

PHOTO : CHART 7 Servcing Expense Ratio

PHOTO : CHART 8 Servicing Productivity

PHOTO : CHART 9 Problem Loan Expenses

PHOTO : CHART 10 Amortized Cost of Purchased Servicing (ACPS)

PHOTO : CHART 11 High Profit-Average Firm Comparison

Robert M. Rosenblatt is an economist with the Mortgage Bankers Association of America in Washington, D.C.
COPYRIGHT 1992 Mortgage Bankers Association of America
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 1992 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Title Annotation:Mortgage Bankers Association of America
Author:Rosenblatt, Robert M.
Publication:Mortgage Banking
Date:Jan 1, 1992
Previous Article:A view of the California market.
Next Article:Saying goodbye to Mr. Birk.

Related Articles
Seeking better industry data.
The MERS project.
A new servicing cost study.
Local programs launched in Birmingham. (Business Alert).
HUD proposes RESPA reforms. (Business Alert).
Senate committee hears industry views on GSE 'bright line'.
Mortgage origination profits plummet in 2004.
MBA paper seeks to clarify roles of mortgage banker, broker.
Strength in numbers.

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