The Peer Group Study.

A key performance study that matches mortgage companies to their peers showed production returns on equity suffered in 1999 compared with the year earlier. But as production performance slipped, average servicing returns rose from 8.1percent to 12 percent for the combined 41 companies in the study.

ED KOCH, FORMER MAYOR OF NEW YORK CITY, USED TO ask his constituents, "How'm I doin'?" A natural corollary to this question is, "Compared with what?" As humans, we often learn about ourselves through comparison with others. Whether being graded on "the curve" or competing on the athletic field, a fundamental way we measure performance is by looking at how we stack up against our peers.

The Peer Group Roundtable began in 1998 as a joint venture between the Mortgage Bankers Association of America (MBA) and The STRATMOR Group. The purpose of the peer group program was to create a means for participating companies to review their results versus peers on a semiannual basis.

Participating companies submit data every six months (June 30 and December 31). The data is compiled and a variety of operating and statistical reports are generated for both servicing and production channels. The companies are divided into distinct peer groups, and senior management representatives (typically chief executive officers, chief operating officers and chief financial officers) from each company meet at an off-site location for a day and a half of roundtable meetings.

Participants drive the meeting agenda.

Historically, participants have discussed overall strategies, reviewed operating tactics and simply compared their results with that of peers. The discussions are usually open and honest. Most firms believe there are very few trade secrets in this business. The way companies differentiate themselves is most often in the execution.

Participating firms certainly benefit from timely, semiannual data comparison. However, most participants tell us they derive at least as much value from the dialogue with peers at the off-site roundtable meetings. Reading a data book is one thing, but only by discussing the results with peers can member firms get "behind the numbers."

Before each reporting cycle, each firm receives a detailed set of definitions and instructions as well as a blank Microsoft(r) Access database. Each of the participating companies gathers the required data and enters it into the database. The file is then e-mailed to MBA, where reports are generated and reviewed in detail by MBA and STRATMOR Group staff, who then prepare a list of follow-up questions for the company.

Even with a detailed set of instructions, we found there are often differing interpretations. Our experience tells us there is no substitute for a thorough review and scrubbing of the data. In fact, this process is often a learning experience for the firms involved--particularly the smaller firms.

This reporting cycle

Forty-one companies participated in the MBA/STRATMOR Peer Group Program for the December 31, 1999, cycle. They were, in alphabetical order, ABN AMRO Mortgage Group; Bank of America Mortgage; Bank of Oklahoma; Bank United; BB&T; B.F. Saul Mortgage Co.; Capital Mortgage Funding, LLC; Carolina First Mortgage Co.; Central Pacific Mortgage Co.; Charter Bank; Charter One Mortgage Co.; Chase Manhattan Mortgage Corporation; Commerce Mortgage Corporation; Continental Capital Corporation (HomeStreet); CrossLand Mortgage Corporation; CUNA Mutual Mortgage Corporation; First Nationwide Mortgage Corporation; Fleet Mortgage Group, Inc.; Fort Worth Mortgage Corporation; GMAC Commercial Holding Corporation; Green-point Mortgage Funding, Inc.; Guild Mortgage Co.; HomeSide Lending, Inc.; Homestead USA, Inc.; Huntington Mortgage Co.; Irwin Mortgage Corporation; M&T Mortgage Corporation; Market Street Mortgage Corporation; MidAmerica Bank, FSB; Mission Hills Mortgage Corporation; Nationwide Home Mortgage Co.; New South Federal Sa vings Bank; Old Kent Mortgage Co.; Pulte Mortgage Corporation; Quicken Loans, Inc.; Roslyn National Mortgage Corporation; Secured Bankers Mortgage Corporation; Temple-Inland Mortgage Corporation; Universal Lending Corporation; Wachovia Mortgage Co.; and Waterfield Mortgage Co., Inc.

We divided the companies into the following peer groups:

* Peer Group A: Large lenders/servicers;

* Peer Group C: Small/medium-sized lenders;

* Peer Group E: Thrift-owned lenders; and

* Peer Group M: Megalenders/servicers.

For the December 31, 1999, meetings, our philosophy was to divide the peer groups generally by annual production and servicing portfolio volume. The lone exception to this was Group E, which is made up of thrift-owned lenders that often engage in portfolio lending. The average characteristics of each group are summarized in Figure 1.

At a glance, one can see the large variety of firms in the program. Each firm receives a data book that summarizes the detailed results by company for its group. In addition, each firm receives summary weighted and simple average results for the other peer groups--so small firms can see the overall results for the larger firms and vice versa. For example, smaller servicers can see how their cost-to-service metrics stack up against larger players. The data includes detailed operating results by production channel. Figure 2 shows the production channel breakdown by group.

A simple review of firm participation in the various production channels can create interesting discussions. For example, one would not expect the smaller firms in Groups C and E to be as active in the broker and correspondent channels. Therefore, why are six out of seven Group E firms participating in the broker channel? (In this context, broker channel means receiving loans from brokers and closing them. It does not refer to taking a retail application and "brokering" the loan to third-party lenders.) Do the results justify their presence in that channel?

Overall results

Is there such a thing as a normal year in mortgage banking? We are a classic cyclical industry and 1999 typified our roller-coaster existence. The first half of the year was a carryover from 1998, the greatest year in mortgage banking history to date. By midyear, however, the party was over as rates increased and the refinance component of production dropped from more than 50 percent to less than 15 percent for many firms.

The increase in rates had another predictable result: Runoff rates declined, servicing values rose and servicing operating costs dropped as portfolios stabilized. Ironically, a great first half of 1999 and a dismal second half may have added up to a normal year in total. However, many firms limped across the finish line on December 31, 1999, with the gloomy prospect of a difficult year in 2000 lying ahead.

One fundamental metric used by many firms is return on equity (ROE). However, the actual equity of participating firms varies greatly. Smaller, private firms may keep only a minimal amount of equity in the company to maximize after-tax cash flow. Larger firms that are part of a bank-owned group may manage equity globally. The equity that "lives" on the books of the mortgage subsidiary may not be meaningful. As a result, our study uses a required equity amount that is calculated using a consistent formula for all participants. The formula considers how much equity a firm would need to operate the servicing and production business on a stand-alone basis, assuming certain leverage factors.

The weighted average return on required equity (RORE) for all peer groups for 1999 is summarized in Figure 3. Acknowledging that the mix of companies is not identical from period to period, the results clearly show the year-to-year drop in production returns from 78.9 percent in 1998 to 52.0 percent in 1999. Also, servicing returns improved from 8.1 percent in 1998 to 12.0 percent in 1999.

Interestingly, the servicing RORE for Group M was the lowest by far, at 10.1 percent. The casual observer may jump to the conclusion that the megalenders' returns lag behind those of smaller servicers. We must be careful here. One of the key factors in computing required equity is the assumed leverage factor. To provide apples-to-apples comparisons, we have used the same leverage factor for all companies in the study. However, while you might give us an A for effort in attempting to create meaningful comparisons, is it really appropriate to use the same leverage factor for servicers of all sizes? To address this issue, we are preparing a sensitivity analysis of RORE, assuming a variety of leverage factors for our June 30, 2000, Peer Group Survey.

A fundamental question we attempt to answer in our peer group program is "Did companies really create economic value during the year, considering the cost of the capital employed to produce the return?" To help us answer this question, we have created a metric called mortgage banking return on capital (MBROC). MBROC considers the total return of the company less the cost of capital employed to generate that return, divided by the calculated required equity. To illustrate, if Company A and Company B produced identical returns but Company B required twice as much capital to produce those returns, Company B would have a larger capital charge and a correspondingly lower MBROC.

The weighted average MBROC for all peer groups for 1999 is also summarized in Figure 3. On an overall basis, production MBROC declined from 42.9 percent in 1998 to 15.9 percent in 1999. Servicing MBROC improved to -8.8 percent from -12.6 percent the previous year. What can we conclude from these results? They suggest that, on an overall basis, the participating firms did not create any shareholder value in 1999 after considering the cost of capital.

But let's remember our earlier discussion regarding leverage. Have we judged the larger firms too harshly by using a leverage ratio that may be too low for them, resulting in a larger required capital and lower RORE and MBROC? This issue has been an excellent discussion topic for participating firms. Regardless of where firms stand on the leverage issue, the dialogue itself has challenged firms to think about their results in the context of economic value added, taking into consideration the cost of capital.

Why is Group M's MBROC so low? First of all, the large volume of lower-margin correspondent business tends to drag down group M's overall margins. One might also make the case that the correspondent channel should have a lower cost of capital than the other channels, given the decreased risk inherent in that channel. However, our survey ascribes the same cost of capital for all production channels. As a result, our one-size-fits-all approach unduly burdens companies in the lower-risk, lower-margin correspondent channel with a heavy capital charge. Going forward, we are considering varying the cost of capital by channel to better reflect risk and return expectations.

What does a mortgage banking firm do for a living? Fundamentally, mortgage firms create loan servicing value. It is hoped that the value of the servicing created is greater than the costs incurred to create that value. One of the "big-picture" metrics used in the MBA/STRATMOR peer group program measures the excess of servicing value over the cost to create servicing rights. The overall results are captured in Figure 4.

Servicing values represent service release premiums received and/or capitalized originated mortgage servicing rights (OMSRs). The cost to create servicing is derived by taking the production income statement and pulling out both the capitalized OMSRs and service release premiums received. The excess value is therefore the pretax net margin of the production business. The data shows that while servicing values capitalized/received increased by 20 basis points in 1999, the cost to create the servicing rights increased by even more basis points). This resulted in a reduction of all-in production margins, from 32 basis points to 13 basis points in 1999.

Because creating servicing rights is what mortgage bankers do for a living, the economic and recorded book value of servicing rights is often the subject of lively discussion at our peer group meetings. If you want to have an interesting discussion (interesting for a mortgage banker, anyway), put a bunch of executives in a room and discuss the economic value of servicing and all the important but subjective factors involved.

So how do peer group firms value OMSRs, and what are the implications? We realize firms may record servicing rights at amounts lower than economic value to be conservative and reduce future earnings volatility. Therefore, we asked member firms to provide us with their best estimate of the economic value of originated servicing rights by product type, regardless of what they record for book purposes. Figure 5 summarizes the results.

In reviewing the value of servicing rights by product type (30-year fixed, 15-year fixed and one-year ARM), Group C and Group A appear to be in the same ballpark. Group M values, however, are consistently 20 to 25 basis points higher across all products. Why are servicing rights apparently worth more to the megalenders? Possible explanations include lower servicing costs, lower guarantee fees on conventional agency product, superior retention and cross-sell capability and a greater ability to hedge servicing rights in an efficient, cost-effective manner.

If servicing values are truly 20-plus basis points greater for the Group M companies, what are the strategic implications? Over time, assuming the servicing values are real and not based upon perceived advantages, Group M companies may be able to pass the revenue advantage through to consumers in the form of lower prices, thus gaining market share.

Production results

Let's shift our focus to the production side. Figure 6 summarizes the average pretax margins by group for each production channel. Overall, you can see that the average pretax margin for all channels combined declined from 32 basis points in 1998 to 13 basis points in 1999. Note that the larger firms in Groups A and M clearly outperformed the smaller Group C and Group E firms. Also note the apparent margin advantage that Groups A and M have in the correspondent channel. Direct lending provided superior margins in 1998 and 1999, but our participants tell us the good times are over for the time being, as portfolio retention units that focused on refinances are sitting on their hands in 2000.

What happened with refinance production from 1998 to 1999? Figure 7 depicts refinance percentages by channel between years. Generally, refinances dropped from 50 percent of production in 1998 to 33 percent in 1999. However, there are some interesting variations by channel. Note the higher refinance rates in the broker and correspondent channels versus retail in both 1998 and 1999.

Given this trend, do firms pay lower service release premiums on loans originating in broker and correspondent channels? Should they? Also, the direct marketing channel moved from almost exclusively refinance business in 1998 (94 percent) to 77 percent in 1999.

What can we learn by comparing retail channel results for small, large and megalenders? Figure 8 provides some interesting insights. Note the range in annual production volumes, from an average of $854 million for Group C to an average of $10.4 billion for Group M. Also note the relatively low government production mix (9.55 percent) for the Group M lenders.

Overall, Group M produced a pretax margin of 47.72 basis points--superior to the other two groups. The primary driver of this is lower production expenses, production support and corporate administration expenses for the Group M companies. The lower production expenses appear to be the result of greater scale, higher refinance mix and a focus on conventional business. Group M companies produced almost 50 percent more loans per production full-time equivalent (FTE) and their volume per branch was more than double the smaller lenders.

On the revenue side, Group M lags behind the smaller companies. The lower revenues are brought about by the lower government loan mix and greater pricing subsidies given to borrowers. The results seem to indicate that the megalenders use their cost advantage to price more aggressively and at the same time maintain superior margins.

Now let's take a look at the broker wholesale channel. Figure 9 summarizes broker channel results for Groups C, A and M. As in the retail channel, Group M has a clear size advantage, a markedly lower government loan mix and a higher percentage of refinance transactions in 1999.

The megalenders enjoy a margin advantage because of significantly lower expenses. Figure 9 shows that Group M expenses are 50 percent lower than the Group C and Group A lenders. The loans closed per production FTE and volume per branch metrics tell the story. This superior productivity and resultant cost advantage in the broker channel drops right to the bottom line for Group M lenders as they registered a weighted average pretax margin of 61.51 basis points.

The retail and broker channel results demonstrate the competitive advantages enjoyed by the megalender group. The advantages include lower direct production costs, lower support and corporate administration costs as a result of much larger scale. The lower costs allow the megalenders to offer lower prices while maintaining superior margins across virtually all channels. These advantages have driven many smaller competitors to origination-only strategies with an emphasis on government production.

In addition to looking at the various production channels, the MBA/STRATMOR peer group program collects and reports data on production support departments such as postclosing, shipping and delivery, secondary marketing, quality control and support/other. The costs per loan are then allocated to the various production channels.

Weighted average production support allocated to the production channels is summarized in Figure 10. The obvious conclusion to be drawn is that the megalender firms in Group M have a clear scale advantage here. By spreading the support costs over a large production base, Group M drives the support costs down to $82 per loan, well below the other groups.

Servicing rights

As we shift the discussion to the servicing side of the shop, let's look first at the big picture. How well did peer group firms do in 1999? Overall servicing results in basis points are summarized in Figure 11. When analyzing servicing results, we look at both net income and net operational income.

Net operational income excludes certain items such as OMSR amortization, impairment/hedging gains or losses and bulk sale gains that more or less tend to be driven by external factors. For the most part, we can see that overall net income improved from 1998, while net operational income was amazingly consistent between years and within peer groups in 1999. There were no surprises in the productivity statistics, as larger firms were clearly superior in loans serviced per FTE.

Weighted average servicing revenues by category (in basis points) are summarized in Figure 12. Most of the differences between groups occur on the OMSR amortization and interest income/expense line items. Are larger firms more likely to be aggressive in capitalizing OMSRs, and therefore are we seeing a correspondingly higher amortization level in servicing? Peer group firms discussed this issue at our meetings last spring. The larger firms were more highly leveraged in servicing, resulting in a negative interest spread of 1.7 basis points.

Servicing-direct expenses (dollars per loan) are summarized in Figure 13. Note the $30 per loan reduction in direct expenses from Group C to Group A. There appears to be significant economies of scale as firms grow from approximately $3 billion to $14 billion in servicing. Interestingly, the megalenders only show an $8 per loan advantage over Group A, yet the average Group M portfolio is about $156 billion. In fairness to the megalenders, some of the expense is affected by the way we allocate corporate administration expenses.

Corporate administration encompasses the following areas: executive, finance and accounting, human resources, information technology and overhead allocations from the parent not directly allocable to production or servicing. The MBA/STRATMOR program summarizes corporate administration costs and allocates to production and servicing based upon a consistent formula for all participants, which enhances the comparability of the data. While participating companies may debate the propriety of the allocation formula, the good news is that all companies are treated the same. We avoid the problem of debating whether Company A has a better allocation methodology than Company B. Corporate administration allocation to production channels in dollars per loan is summarized in Figure 14. Consistent with production support, the scale advantages of the larger firms is borne out by the results.

Do any of the firms appear top-heavy? To help us address this issue, corporate administration and technology employees as a percentage of total employees are summarized in Figure 15. Group M appears to have a high percentage of administrative employees, but this is partially due to its investment in technology. The apparent payback for Group M companies comes in the form of superior margins, particularly in the broker and correspondent production channels. Group A appears to be light in administration employees, but many of the Group A firms are bank-owned. Some Group A firms received allocations of parent costs but without a corresponding allocation of FTEs.

Future plans

Where are we going with the peer group program? On the technology front, plans are being made to enhance data collection and reporting capabilities for the peer group program. A Web interface is being built to enable participating companies to submit data over the Internet with built-in edit checks at the point of entry. Ultimately, companies will be able to generate their own reports via the Web.

One of the challenges of a peer group program is that the mix of companies is ever-changing, as firms are bought and sold and new firms join the program. Year-to-year comparisons become more difficult. Our new reporting tools will allow us to run reports that will show comparisons between a fixed population of companies from period to period.

Thirty-six firms participated in the December 31, 1998, roundtables and 41 firms in December 31, 1999. We expect approximately 45 firms to participate in the MBA/STRATMOR peer group program for the June 30, 2000, cycle. In addition to expanding the current groups described above, we are actively pursuing subprime firms, a special servicer group and a community bank group.

Historically, the MBA/STRATMOR peer group program's focus has been strategic, with much of the value derived from the interaction of chief executive officers and chief financial officers at the roundtable meetings. In addition to the peer group program, the MBA sponsors other surveys such as the Cost of Servicing Study and the IT Cost Study. Fundamentally, these surveys are more detailed than the peer group program and have an operational emphasis.

Because of the multitude of surveys, our participants sometimes feel "surveyed out." To address this, in the short run we have attempted to make definitions consistent across the various MBA surveys. In the long run, we envision a day when we combine the survey data-collection process using a Web interface. Then a variety of meetings could be held that focus on certain aspects of the data, such as servicing and information technology.

The MBA/STRATMOR peer group program has been popular with participants because they get something out of it. The value starts with the quality and timeliness of the data, but it does not end there. Participating member firms tell us they leave our meetings with new insights on strategic trends in the industry and a keen awareness of the issues that need to be addressed at their companies. We look forward to the future as we attempt to keep up with the rapid pace of change in our industry.

Jim Cameron is an Atlanta-based partner with The STRATMOR Group.

Figure 1
Average Characteristics of Peer Groups
                     Group A                 Group C
Millions    (large lenders/servicers)  (Small/medium-sized
                                            (lenders)
Assets                  874                     326
Production            5,044                   1,006
Servicing            13,859                   3,501
                   Group E                  Group M
Millions    (thrift-owned lenders)  (megalenders/servicers)
Assets              1,120                     7,537
Production          1,385                    43,468
Servicing           4,495                   155,964
Millions    1999 Total
Assets         1,924
Production     9,951
Servicing     40,148
SOURCES: STRATMOR GROUP MBA
Figure 2
Production Channel Breakdown by Peer Group
                 Group A  Group C  Group E  Group M
Retail             10       13        7        6
Broker              9        5        6        6
Correspondent       7        3        4        7
Direct              9        4        2        7
Subprime            4        0        1        1
Total Companies    12       15        7        7
SOURCES: STRATMOR GROUP, MBA
Figure 3
                           Group A  Group C  Group E  Group M
Weighted Average RORE for
all Peer Groups, 1999
Production                  35.6     36.7     -28.1     57.7
Servicing                   23.8     23.2      53.9     10.1
Overall                     26.9     27.2      31.4     17.7
Weighted Average MBROC
for all Peer Groups, 1999
Production                  -0.4      0.6     -64.2     21.7
Servicing                    3.0      2.4      33.2    -10.7
Overall                      2.1      1.9       6.4     -5.5
                           1999 Total  1998 Total
Weighted Average RORE for
all Peer Groups, 1999
Production                    52.0        78.9
Servicing                     12.0         8.1
Overall                       18.9        33.6
Weighted Average MBROC
for all Peer Groups, 1999
Production                    15.9        42.9
Servicing                     -8.8       -12.6
Overall                       -4.5         7.4
SOURCES: STRATMOR GROUP, MBA
Figure 4
Excess of Servicing Value Over the Cost to
Create Servicing Rights
                          1999 Total  1998 Total
Servicing Values             152         132
Cost to Create Servicing     139         100
Excess                        13          32
SOURCES: STRATMOR GROUP, MBA
Figure 5
Economic Value of Originated Servicing Rights, by Product Type
                          Group C  Group A  Group M
30-Year Fixed Conforming   1.47     1.45     1.64
30-Year Fixed Jumbo        1.20     1.09     1.36
30-Year Fixed Government   2.08     2.12     2.38
15-Year Fixed Conforming   1.13     1.18     1.38
15-Year Fixed Jumbo        1.04     0.93     1.18
15-Year Fixed Government   1.50     1.71     1.94
l-Year ARM Conforming      1.07     0.99     1.11
1-Year ARM Government      1.36     1.51     1.75
SOURCES: STRATMOR GROUP, MBA
Figure 6
Average Pretax Margins by Peer Group for each Production Channel
               Group A  Group C  Group E  Group M  1999 Total
Retail            30        9      -22      41         15
Broker             2        1        4      49         14
Correspondent     28      -53      -71      20         -5
Direct           119       -2       28      88         77
Subprime          62      N/A      -32     -91         21
Average           27        6      -18      32         13
               1999 Total
Retail             46
Broker             25
Correspondent     -10
Direct            140
Subprime          114
Average            32
SOURCES: STRATMOR GROUP, MBA
Figure 7
Refinance Percentages by
Channel, 1998-1999
                 1999                 1998
               Purchase  Refinance  Purchase  Refinance
Retail            67        33         50        50
Broker            59        41         41        59
Correspondent     60        40         45        55
Direct            23        77          6        94
Subprime          26        74         45        55
SOURCE: STRATMOR GROUP, MBA
Figure 8
Comparisons of Retail
Channel Results for Small, Large
and Megalenders
                                Group C  Group A  Group M
Number of Companies               13       10        7
Origination Volume ($millions)    854     2,302   10,425
Government Percentage            25.27    33.20    9.55
Refinance Percentage             35.47    26.78    45.58
Weighted Average P&L (bps):
Income                          269.82   295.87   229.02
Production Expense              190.27   228.02   160.76
Allocated Production Support     16.96    19.01    7.66
Corporate Admin. Allocation      39.12    19.51    12.88
Net Income                       23.48    29.34    47.72
Loans/Production FTE              29       32       43
Volume per Branch               18,089   34,516   76,004
SOURCE: STRATMOR GROUP, MBA
Figure 9
Broker Channel Results for
Groups C, A and M
                                Group C  Group A  Group M
Number of Companies                5        9        6
Origination Volume ($millions)    405     2,800   10,587
Government Percentage            42.0     30.7     12.3
Refinance Percentage             34.0     42.9     49.3
Weighted Average P&L (bps):
Income                          131.68   111.67   115.19
Production Expense               59.21    78.76    39.65
Allocated Production Support     25.62    13.09    4.84
Corporate Admin.Allocation       25.91    14.26    9.19
Net Income                       20.94    5.56     61.51
Loans/Production FTE              143      105      171
Volume per Branch                155.6    320.0    599.3
SOURCE: STRATMOR GROUP, MBA
Figure 10
Weighted Average
Production Support Allocated to the
Production Channels
              Group A  Group C  Group E  Group M  1999 Total
$/Loan          190      222      379      82        110
Basis Points     16       19       26       6          9
SOURCE: STRATMOR GROUP, MBA
Figure 11
Overall Servicing Results,
in Basis Points
                        Group A  Group C  Group E  Group M  1999 Total
Net Income                  13     12        24         7        13
Net Operational Income      27     24        29        29        27
Loans Serviced Per FTE   1,154    839       898     1,211     1,030
                        1998 Total
Net Income                   6
Net Operational Income      26
Loans Serviced Per FTE     955
SOURCES: STRATMOR GROUP, MBA
Figure 12
Weighted Average Servicing
Revenues by Category,
in Basis Points
                                 Group A  Group C  Group E  Group M
Service Fees/Late Charges/Other   33.8     36.9     39.2     40.0
Interest Income/Expense            4.0      3.8      4.5     -1.7
OMSR Amortization                -18.2    -14.9     -9.2    -23.4
Impairment/Hedge Gain/Loss         3.2      1.4      1.3      1.1
Bulk Sale Gains                    2.2      1.3      2.6      0.7
Total                             25.0     28.4     38.3     16.7
                                 1999 Total
Service Fees/Late Charges/Other     39.2
Interest Income/Expense             -0.8
OMSR Amortization                  -22.4
Impairment/Hedge Gain/Loss           1.4
Bulk Sale Gains                      0.9
Total                               18.3
SOURCES: STRATMOR GROUP, MBA
Figure 13
Servicing-Direct ExpensesServicing-Direct Expenses
(dollars per loan)
                                    Group A  Group C  Group E  Group M
Average Portfolio ($billions)        13.9      3.1      4.5     156.0
Simple Average Direct Expense/Loan     70      100       84        62
                                    1999 Total
Average Portfolio ($billions)          40.1
Simple Average Direct Expense/Loan       79
SOURCES: STRATMOR GROUP, MBA
Figure 14
Corporate Administration Allocation
Allocation to Production Channels
(dollars per loan)
                 Group A  Group C  Group E  Group M  1999 Total
Channel Average    217      398      394      192       312
SOURCES: STRATMOR GROUP, MBA
Figure 15
Corporate Administration and
Technology Employees as a
Percentage of Total Employees
                                 Group A  Group C  Group E  Group M
Technology Percentage              2.4      3.2      4.8      6.1
Other Administrative Percentage    4.2      6.1      7.9      5.0
Total Administrative Percentage    6.6      9.3     12.7     11.1
                                 1999 Total
Technology Percentage               4.9
Other Administrative Percentage     5.0
Total Administrative Percentage     9.9
SOURCES: STRATMOR GROUP, MBA