The demands of customers and the evolution of the financial marketplace dictate that banks view their role differently.
To recapture market share, differentiate themselves, become more efficient, and manage their risks, banks will have to enhance their expertise - in other words, become expert banks.
In the first article in this series I looked at the reasons why banks need to become expert. Now I'll discuss how that can be accomplished.
The first step is understanding the kinds of expertise banks need to accumulate and use. In general, there are three kinds of expertise pertinent to financial services: rules, associations, and models.
Rules are just what you would think they are, and come from a variety of sources. Internal rules, sometimes called policies, are set out by the financial institution itself, and can cover such things as position limits, underwriting requirements, and processes. External rules can come from the law of the land, industry regulators, accounting practices, or agreements the institution has entered into.
Associations, sometimes referred to as artificial intelligence, connect one event or rule to another. For example, if a customer lives in a certain ZIP code and bought a house 10 or more years ago, association would tell us that person might be a good candidate for a home equity line of credit.
Similarly, if a corporate customer submits a letter of credit for collection in Asia in 90 days, association would tell us that company might be a candidate for a forward currency transaction and/or 90-day receivables financing. The difference between rules and associations is that rules are restrictive and associations are creative.
Models are a combination of rules and associations that represent the behavior of something or someone under a set of conditions. Models might tell us that if the GNP grows by less than X% over the fourth quarter, Y% of our credit card balances will become delinquent, or, if we increase the management fee on our 401(k) accounts by a certain amount, a certain percentage of them will shift to our competitors.
Though models are attractive as decision support aids, they are only tools and should not be mistaken for reality.
In fact, when reality and the model diverge, it is often a good sign that trouble is brewing.
These three kinds of expertise apply to all businesses, but within them there are classes of expertise that apply specifically to financial institutions.
One kind of expertise that banks should pay particular attention to is suitability in the sale of investment products. Securities firms are aware that the securities regulators and the courts will hold the selling institution responsible if a customer can prove that an investment was unsuitable for the buyer at the time the purchase was made.
Although banks that sell investment products generally make sure their salespeople have the right registrations, they are much less vigilant about whether those salespeople ensure that each investment is suitable for each customer.
Another specialized kind of financial expertise relates to the price performance of complex instruments, particularly derivatives, which come in all flavors.
There has been a lot of press coverage recently about derivatives losses in organizations like Barings and Daiwa, but in fact the strategies and instruments used in both of those cases were not overly complex.
The expertise necessary to prevent these problems is not on the same level as that needed to prevent the kind of losses, for example, that are now being adjudicated between Bankers Trust and Procter & Gamble.
Once we understand the nature of the necessary expertise, we need to know how to accumulate it.
"How to" really falls into two categories: identifying the expertise you want to accumulate and deciding the receptacle in which it should be accumulated.
Identifying the expertise might appear straightforward, but it often is not.
Some rules are fairly straightforward, particularly if they are well codified, and those rules can be accumulated from the manuals.
On the other hand, there are many times when the rules by which processes actually get done and decisions actually get made are not what appear in the manuals at all.
In this case, accumulating the rules from the manual can be dangerous, and management must be careful to determine what expertise is being accumulated.
Accumulating associations is somewhat different. While rules tend to leave relatively little to interpretation, many associations are nothing but interpretation. While rules are often codified in detail, associations are seldom written down anywhere. Rules are frequently made clear at the start, but associations may not become clear until relatively late in the game. And, unfortunately, computers tend to be unsuited to, and unhelpful at, identifying associations.
This last fact goes a long way toward explaining the importance of large-scale parallel processing in the development of decision support systems, since the associations necessary to decision support are often apparent only after a tremendous amount of experience.
Models are made up of both rules and associations, so they have two aspects with regard to identification. Because models are artificial representations, they are often constructed instead of identified.
The problem comes in assessing and correcting the model over time, so that it moves ever closer to reality. Since assessment and correction make up by far the harder part, many model builders neglect it, and then wonder why their models led them astray.
All this accumulating has no value unless the expertise can be accessed and used, so we must pay special attention to the receptacle for the expertise. In most cases, the purpose of accumulating expertise is to enhance decision-making, so the storage format must adapt itself to that use.
For many rules, and some associations, a data base with a decision tree structure or facility is the best choice. Then, once users accesses the expertise they need, they can follow the tree to the correct conclusion.
Models, because of their special nature, also require a flexible receptacle, capable of accepting additional expertise and running the model itself.
Since models can be quite complicated, building a receptacle that can accept the data necessary to run the model, perform the necessary calculations (sometimes in reverse), and then allow the experts to modify the model without affecting its quality is a daunting task.
People who have the requisite understanding of both the banking business and the workings of models are extremely hard to find. Banks that already employ such people should make it worth their while to stay - and then make sure they are getting the most out of their talent.
Like the Library of Congress without a card catalogue, a bank with lots of expertise but no efficient access is an expensive white elephant. And the similarity doesn't stop there. Just as the way people access books is changing, the way bankers need to access expertise is changing. Just as the Library of Congress must make many of its books available in new ways, banks must make expertise available in new ways.
The access method most commonly thought of is "by request," where users indicate to the system that they want access to a certain expertise. In many cases this access method is preceded by a search, since the employee cannot be expected to know all the expertise that the bank has accumulated. Searches can be run by subject, text string, or other categories.
Of much more value to both the bank and the employee is a "triggered" access, in which an action by an employee triggers an automatic access to, and presentation of, the expertise.
In this case, the calling up of the expertise automatically becomes part of the process. This kind of access can be nonintrusive, where the expertise is made available to the employee, or intrusive, where the process cannot continue until the employee follows the rule or activates the model.
As customers perform more transactions for themselves, they have more need for expertise, and banks will do well to afford them ever wider access to it. Discount stockbrokers have already begun to offer their customers on-line access to research and recommendations, and Intuit has led the way in providing the household with some financial expertise.
Many banks are rushing to establish home pages on the Internet, but that by itself provides little value. The banks that allow their customers to get at some of their financial expertise through the Internet may finally turn that network into a marketing medium instead of a promise.
Accumulation and access do not come free - in fact, they cost a fair amount of money. In a time when banks are under constant pressure to reduce costs, how can management justify the expense of accumulating and accessing expertise?
The answer to this question is at the heart of banking's most difficult challenge - looking beyond a simple infatuation with cost reduction.
My research indicates that there is no beneficial impact from the simple reduction of noninterest expenses, either in the United States or around the world.
Part of the reason for this is that some banks understand how to value expertise, and their higher expenses can be turned into higher profits.
The simplest way to value expertise is to look at it as an investment, just as you would a computer or an office building. In these cases, one either looks at the cost of the investment and compares it with the present value of the expected income improvement, or one looks at the amortization of the investment versus the projected flow of benefits.
In either case, the analysis generates a "hurdle rate" or "hurdle present value" below which the investment is not made.
The problem with such analyses is that, unlike many other investments, the value of expertise increases the more it is used.
Thus the valuation is heavily dependent on the enthusiasm and ingenuity of the bank's employees who must use the expertise. For this reason, the employees must be involved in the valuation, and must have an incentive to use the expertise to create value.
One way to accomplish this is to involve the employees' compensation system in the cost and utilization of expertise. For example, charging some portion of the development and carrying costs against employees' profit- based bonus pools will give management a good reading as to whether employees believe in the investment, and it will motivate the employees to make the most of the bank's investment.
In the end, no matter how well designed the expert system might be, or how well the incentive program is constructed, turning expertise into profits is up to the people who work at the bank - the bankers themselves.
If they are driven by a constant need to improve themselves, if they measure their expertise against their peers on a regular basis, and if they constantly solicit input from both management and customers, they are likely to make good use of the bank's expertise. At the same time, they are likely to add to the accumulated store of knowledge.
It is this giving and taking of the wealth of expertise that marks a winning institution and winning individuals.
Treating knowledge like a crop to be hoarded for a rainy day may be a natural reaction, but it is the worst kind of reaction. What is needed is a partnership between the bank and its people, where both prosper by contributing and taking at the same time. It is truly a cultural thing, and that kind of culture breeds both expert banks and expert bankers.
Mr. Bollenbacher is director of strategy and business development for the worldwide finance industry division of Unisys Corp., Blue Bell, Pa. The first article in this series appeared Dec. 28.
