Automation can go too far in lending
* * *
Just before Memorial Day there was considerable attention in the media to the discovery and the subsequent recovery of information from the "black box" containing flight data from the final moments of the Air France Flight 447. This was an A330 Airbus, one of the most sophisticated civilian aircraft in the world that went down in a trans-Atlantic flight from Rio de Janeiro to Paris two summers ago with the loss of all 228 aboard.
Even after two years submerged nearly two miles below the surface, the flight recorder information was largely intact. It appears to reflect the actions of a crew that was confused by contradictory information. The crew did not perform certain standard maneuvers, such as when a plane loses air speed, to push the nose down to increase air speed and lift.
The tentative conclusion of the investigators is that the crew (and by extension, most all air crews of civilian airliners) may be insufficiently drilled on the basics. They apparently rely too heavily on information that is computerized and automated and subject to mechanical error or, at a minimum, subject to misinterpretation due to conflicting information and signals.
Automation makes sense ...
When I read the comments about automation, I immediately thought of the trend in business lending in recent years toward automating as much of the loan decisionmaking as possible. This has been going on now for many years and to a considerable degree makes financial sense.
First, staff salaries are the second-largest item of expense for banks ranking behind interest costs to fund loans.
Any means to limit these costs or to curtail their growth are understandable and worthy objectives of any management group seeking to legitimately maximize return on stockholders' equity.
Second, lending relies on a series of decisionmaking metrics, some of which are relatively easy to automate.
These include the calculation of a handful of financial ratios that give relatively reliable indications of the borrower's liquidity, solvency, and profitability.
Third, most small business lending relies on the demonstrated payment history of the principals of the business.
These are conveniently and reliably summarized as scores in credit bureau reports. While these are bureau scores of individuals, at the smaller end of the scale, the "signature" of the business from that of the owner is indistinguishable.
Fourth, consistency and reliability of the application of business processes produces a confidence in the outcome of the process.
Enhanced compliance is also an important potential by-product of central underwriting.
...but there are limits
Clearly there are limits to the application of highly automated systems to business credit. Complex transactions require customized and labor-intensive scrutiny and crafting and cannot reasonably be left to black box decisions.
But the volume of small commercial lending is potentially very large and continues to grow at least theoretically, even though business lending activity continues to languish in the current economic environment.
There are longer-term issues that should be addressed and considered by commercial lending managements. Trends toward centralization in the interests of economy and consistency must be balanced by considerations of access to commercial credit by borrowers.
Consider this: In recent years the delivery of small business credit has increasingly become a retail product by delivery and access channel. How many banks are deliberately staffing their branches (or at least those branches that traditionally have been access points for commercial lending opportunities) with people trained to identify commercial credit opportunities and offer basic business advice and counsel?
Does efficiency cost opportunity?
When I arrived at First National Bank in Albuquerque in the 90s, there were among the bank's 30 or so branches at that time about half a dozen where commercial lending opportunities were abundant.
We deliberately staffed those branches with experienced small business lenders as managers or assistant managers. The customers had access to bankers who understood their needs and the bank had representatives in the field that could spot commercial lending opportunities as well as nurture the small business owners and borrowers.
Today, it's relatively rare that branch managers have extensive lending experience. Many community banks have not abandoned the traditional model but many more have as they have grown or been acquired and become much larger institutions.
My question is this: How do community and small regional banks assure a constant stream of small business lending opportunities through their branch networks? The "new" model uses the branches to deliver credit but how does the bank identify lending opportunities?
One method is the "hunter-and-skinner" approach-also referred to, disparagingly, as the "empty suit" model. The hunter is the person with the big broom sweeping the prospects toward the front door of the branch, where the application is taken (perhaps by machine or voice recognition software) and transmitted downtown to the central underwriting unit.
Does anyone seriously think that the customer enjoys this sort of interface? Does anyone think that this develops a customer in an appropriate and helpful way, longer term?
Don't put your brains in a black box
This brings me back to the thought I had when I read about the tentative conclusions of the investigators of the Air France disaster. Companies that rely on black boxes risk creating gaps in what their staffs know and can do. The results of centralized underwriting will not have the tragic results of Flight 447.
But are the best interests of the borrowers and the local community served by impersonal "black box" thinking and practice?
There may not be many long-term alternatives to this trend but to see the traditional process upended in ways that do not benefit the individual borrower, the bank, or the local community is to witness a loss of valuable community resources.
It's another casualty of the sort of efficiency where the pencil may be too sharp to the point where it rips the paper.
Do you agree with Ed? Or do you think the black box is the way to go? Respond in the comment box below.
About Ed O'Leary:
Veteran lender and workout expert O'Leary spent more than 40 years in bank commercial credit and related functions, working with both major banks as well as community banking institutions. He earned his workout spurs in the dark days of the 1980s and early 1990s in both oil patch and commercial real estate lending.
O'Leary began his banking career at The Bank of New York in 1964, and worked at banks in Florida, Texas, Oklahoma, and New Mexico. He served as a faculty member and thesis advisor at ABA's Stonier Graduate School of Banking for more than two decades, and served as long as a faculty member for ABA's undergraduate and graduate commercial lending schools.
Today he works as a consultant and expert witness, and serves as instructor for ABA e-learning courses and has been a frequent speaker in ABA's Bank Director Telephone Briefing series. You can hear free audio interviews with Ed about workouts here. You can e-mail him at
. O'Leary's website can be found at www.etoleary.com.
You can get word about these columns the week they are posted by subscribing to ABA Banking Journal Editors Report e-letter. It's free and takes only a minute to sign up for. Click here.