We have again partnered with The Kafafian Group (TKG) to review and update our default assumptions for loan origination and servicing costs, deposit net operating costs and other deposit parameters. Using TKG’s 2015 industry analysis as the basis, we have developed an updated recommended set of default values which we are sharing.
It should be noted that each bank or credit union must take specific actions to adopt and implement these new default assumptions. They do not happen automatically, and no assumptions will change without a client’s prior approval.
What is not changing?
The article does not address any other assumptions beyond those described herein, including credit capital and annual loss inputs, collateral and guarantee recovery rates, funding costs or Target ROEs. Any questions or assistance needed on other key assumptions should be directed to your Client Success Manager (CSM).
Using your own data
Some of our clients have developed their own overhead cost assumptions based on internal studies. We continue to believe this represents the best information to use in the PL solution. While the default assumptions are reasonable, they do not represent any one institution or financial institution type or size. If your bank has information that reflects the institution’s internal costs, please continue to use that data.
The TKG Sample Banks
While the TKG data from this year’s report included about the same number of institutions (30) as last year, there were some changes in the mix of individual institutions within the sample. A few of their clients left the sample, due to mergers and other reasons, and a few new (to TKG) banks were added. The size of the banks in the sample generally range from $800 million to $2.0 billion in total assets. While the sample included banks from around the U.S., the majority were based in suburban markets in the mid-Atlantic and Northeast regions. The sample banks have a median efficiency ratio of about 67.4%.
Loan Origination and Servicing Channels
As we discussed in last year’s article, the cost to originate a loan includes both direct (loan officers salaries, benefits for credit analysts, etc.) and indirect (accounting expenses, electric utility charges and general marketing, etc.) costs. When determining the appropriate costs to allocate to loans and deposits, we continue to believe that using the marginal costs method vs. the “fully absorbed” costs method is the preferred practice. Loan and deposit pricing is based on making the right marginal economic decision for the next loan or deposit.
We will provide, however, both the PL recommendation that uses the marginal cost method as well as the TKG fully-absorbed median for the clients. To determine the PL defaults we start with the TKG data.
The information they provided includes estimates for various general overhead expenses, including the senior executive group’s total compensation, general marketing expenses and various accounting functions like regulatory reporting. This figure was about 16.7% of total cost, which is similar to last year’s level of 16.4%.
There were some further breakdown of direct and indirect costs. However, what is direct and indirect can vary by institution. We continued the procedure we instituted last year of using the simplified assumption on loan origination that about 50% of the total cost is marginal direct cost. However, on the servicing side we continue to only remove the general overhead expense. Thus, the PL recommendations are modifications of the TKG figures (Modified TKG).
The most recent report indicates greater loan activity in 2015 than in 2014. This can affect the indirect cost figures. As such, we decided to use an average of the 2014 and 2015 median data for each loan type. As also noted last year, the TKG information is for a “typical” loan which tends to be in the $500,000 range on the commercial side. The PL defaults provide further breakdowns, with a lower cost for “smaller sized” loans and higher levels for bigger loans. In addition, for larger community banks we increased the number of higher dollar tiers. The percentage difference in the amounts between tiers in most cases is similar to those previously recommended.
In general, the 2015 cost levels from TKG were slightly less than the 2014 numbers. This was partially due to the increased loan volume driving down “per loan” allocation of indirect costs, which tend to more fixed in nature. Since we used an average of the past two years, this year’s figures are also somewhat lower in most categories than the previous year. It should be noted that this year’s defaults are still considerably higher than those used prior to 2015. The chart that follows shows the TKG levels for 2014 and 2015. As noted, this is on a “fully loaded” basis.
Finally, one frequently asked question is whether the cost structure of TKG’s sample banks could different significantly from the cost structure of our clients. Since this is aggregate data, it is unlikely that any one institution would have the exact costs as shown above. For some of our clients the costs are likely higher because of their operations and locations (e.g. higher salary and rent structures). But for some it is lower.
If your institution is a small bank located in rural Kansas with a 50% efficiency ratio, it is likely your costs are lower than those shown above. However, if your institution is a larger bank in the San Francisco area with an efficiency ratio above 70%, your costs are likely higher. As such, while you should view our recommended default values as directionally correct and reasonable, we would encourage you to consider making any qualitative adjustments to the defaults if your institution’s profile (e.g. size, location, efficiency) is significantly different than the sample banks.
Changes in Channel Structure
PL has traditionally had three major channels of costs per loan type: simple, average and complex. The general view is that most loans would fit in the average category. Those that might be part of a multiple facility or a renewal of an existing loan might be in the simple category. Loans with complicated features or borrowers (like SBA or participations sold) would be complex.
What we are finding is that very few loans, except for some construction type loans, are given a complex cost for either origination or servicing. A brief survey of our clients show generally less than 1% of the opportunities are given a complex designation. The highest we have seen is 8% of the opportunities listed as complex on the origination side and 5% on the servicing side. As expected, most loans are rated average on origination and servicing, with about 15-20% as simple.
Furthermore, we have had questions arise from time to time about how to best define “average” vs. “simple” vs. “complex.” These descriptions are very general and a bit ambiguous. As such, we decided to introduce a new naming convention and method for the channels.
On the origination side, we used a designation of “New Business” for what had “Average” and “Renewal or Multifacility” for “Simple.” The idea is that if a loan represents a new opportunity, the “New Business” costs should be selected, and if the opportunity is a renewal of an existing loan or the second or third loan facility in an opportunity, the lower cost “Renewal or Multifacility” category should be selected.
When a loan opportunity is complex, with the likelihood of longer time spent by the lender and/or bank than the default figures shown, extra costs can be manually added to the channel selected.
Finally, where a loan type is being used as a miniperm or permanent loan to a construction or non-revolving line of credit (e.g. conversion product), we have added the “Conversion” category. This generally has a very limited cost basis and reflects the likelihood that most of the origination costs are accounted for in the initial loan facility.
On the servicing channel, we are also recommending two categories for commercial loans. “Standard” replaces “Average” and “Simple or Multifacility” will be used instead of “Simple.”
For construction loan types, “Simple” will be replaced by “Master Note.” If the loan is more complex, we recommend manually adding extra cost to the category.
For consumer type loans, we are also reducing to two categories in each channel – using “Renewal” and “New Business” for origination and “Standard” and “Simple” for servicing.
Please note: if your bank currently uses the Relationship Awareness module please inform your assigned CSM, consultant or RA representative before making any changes in structure.
Other differences from last year:
- When we increased the costs last year to reflect the TKG figures, we maintained the same percentage differences between the cost for larger and smaller loans. This resulted in lower costs for smaller dollar loans, but still substantially above what had been there in the past. The new levels made it much more difficult to hit ROE targets for the lower dollar loan opportunities and in some cases resulted in negative returns at most reasonable interest rates. We are using a lower percentage of the modified TKG figure this year for the small dollar loans, which should make obtaining a reasonable ROE more attainable.
- In the case of lines of credit, we analyzed the probability of a line to a new customer that is expected to renew in the future and separately the likelihood of an existing line renewing. We used a 50% probability on a new line and a 75% factor on an existing line. This led to a 50% reduction from the modified TKG for new business and about a 75% reduction for existing lines. This should also result in a higher line of credit ROE, compared to the current values.
- We also suggest that the values recommended in item 2 should be used for a Letter of Credit opportunity type. For most letters of credit, the Usage Given Default is 100%, but the probability of a draw down is very low, typically 5% or less. We would recommend for this opportunity type to use a 5% UGD compared to the 35% to 100% associated with lines of credits (Construction loans should remain at or close to 0%). For letters of credit, we suggest UGD should be considered Usage Given Draw.
- We changed the dollar break points from last years – $50,000, $200,000, $600,000 and $1 million – to $150,000, $500,000, $750,000 and $1 million. This reflects some increased authority levels we are seeing at some of our clients.
- We examined the costs this year compared to last year on deposit products. There was an average rise in cost of about 2% in the various categories, although some varied by greater levels. As such, we increased the deposit cost by 2%.
- On transaction deposits, such as DDA and NOW, we had used a conservative default 6% float and reserve figure. However, all of the individual studies we have done for clients indicated a figure of less than 4%. We have decided to recommend a 4% figure this year. We would urge our clients to contact us so that we can do a relatively simple study to determine the exact figure for your institution. We have not made any changes to the other deposit types on float and reserve.
We would suggest a few changes, if not already instituted, in the Reporting Regions. The interpolation method for those using the option approach for caps and floors should be “Monotone Convex.” Also, on the Funding Cost Option we recommend “Do not Include Capital in Funding.” Finally, the Unused Line Opportunity Cost is recommended to be a six-month Transfer Duration and 20% Liquidity Factor.
The appendices include the results from the TGK study and PL’s new cost recommendations for loans and deposits. If you have any questions or comment please contact us at 980-297-7100 or support.precisionlender.com. Download the appendices.