This article is the second of a two-part series on the high-level results of KPMG Peat Marwick's second annual benchmarking survey for subprime mortgage originators and servicers.

The results were based on data from 14 lenders that originate a total of over $14 billion of subprime mortgage loans, and eight companies that service a total of 300,000 loans with total principal balances exceeding $2 trillion.

Subprime mortgage servicing is a high-cost venture, when compared to the servicing of conforming loans, mostly due to the additional resources required in default management.

Among servicers who participated in the survey, the average portfolio size, including loans serviced for others, was 40,000 loans, with an average balance of about $60,000.

To compare the effect of portfolio size, we divided our servicers into two size groups, with 20,000 loans our dividing point. We found that direct costs averaged about $200 per loan, with the larger servicers having costs one-third lower than their smaller counterparts'.

The largest area of difference between the small and large subprime servicers was in customer service, where costs were almost three times greater for the smaller servicers.

The direct cost of $200 per subprime loan is in comparison to direct costs for conforming mortgage servicers of less than $50 per loan.

As expected, the largest difference between the subprime and conforming groups is in default management (collections, workouts, foreclosure, bankruptcy, and foreclosed real estate). There, subprime default management costs seven and one-half times the cost for conforming servicers. The costs in this area are largely due to the lower credit quality of the loans being serviced.

The overall delinquency rate also reflects the lower credit quality. The subprime group averaged 12%, while the conforming group averaged less than 4%.

Economies of scale, however, may also play a role. The average portfolio size for conforming servicers was about 400,000 loans-10 times that of the subprime group.

Indirect costs, which include general and administrative costs (finance, accounting, executive office human resources, legal counsel, etc.), data processing, and record retention, averaged $60 per loan, approximately one- third more than for the conforming group. The average for data processing was $22 per loan versus $18 per loan for the conforming group.

Considering the large difference in portfolio size and the economies of scale that one might expect in technology, the quantitative results coupled with our on-site visits indicate that most subprime servicers are underspending in technology.

For example, only 50% of the subprime participants used a voice-response unit. More spending in technology might result in reduced costs throughout the operation.

Analyzing productivity statistics yields some interesting conclusions as well. KPMG defines direct productivity as total loans in the portfolio divided by the number of direct full-time employees or equivalents. Direct productivity averaged 425 loans per full-time equivalent in the subprime group but almost 1,100 for conforming lenders.

Direct costs and productivity are usually inversely related at equal rates. However, direct costs are four times higher in the subprime survey, yet productivity is only two and one-half times more for the conforming servicers.

Is this due to higher costs per full-time equivalent, a higher percentage of nonpersonnel costs, reduced use of outsourcing, or other factors?

Personnel costs as a percentage of direct costs were the same in both groups. However, personnel costs (including salary, bonus, benefits) per full-time equivalent for the subprime group averaged $44,000 but only $33,000 for conforming servicers. So though they hold fewer loans, the subprime entities paid more per full-time equivalent-in part because of more specialized skills required to handle their customer base.

Revenues were examined in our survey as well, to determine net profitability.

Servicing fees, including fees earned via subservicing less the amortization of mortgage servicing rights, averaged $228 per loan, or 36 basis points. Ancillary income such as late charges, prepayment penalties, and insurance averaged an additional 18 basis points, plus 5 earned as float income, for a total of 59 basis points.

Servicing fees differed by $94 per loan, or 22.5 basis points, compared with conforming servicing fees. Ancillary income was 80% higher for subprime servicers, mostly due to late charges and prepayment penalties.

Thus it appears that subprime servicing is a profitable venture for our participants.However, most recognize that ongoing profitability in servicing is based on managing the cost to acquire or originate loans and on keeping control of delinquencies and losses.

Subscribe Now

Access to authoritative analysis and perspective and our data-driven report series.

14-Day Free Trial

No credit card required. Complete access to articles, breaking news and industry data.