People often ask me, "Is Microsoft a threat to the banks?" This is a very surprising question.

GE Capital has almost $300 billion of financial assets. GMAC has $12 billion of financial services revenue, more than Microsoft's total corporate revenue. Microsoft has no leasing subsidiary, takes no deposits, makes no loans, and offers not a single financial product-in an age when everybody has financial products. Yet Microsoft is viewed as the threat, not GE Capital or GMAC.

People feel this way for a simple reason: Because of advances in information technology, the banking industry is becoming less vertically integrated over time.

During dis-integration, the industry changes. More of the costs and the work are done by nonbanks, which tend to be technology or services companies. Job growth and much of the innovation occur outside the banks, not within. Bankers see this trend as a threat.

Vertical integration is the degree to which an industry or company creates all parts of its product. Does an auto manufacturer make its own steel or does it buy the steel? Does a computer company make its own chips or buy them elsewhere?

It also means the degree to which a company distributes and sells its products itself. Does the automaker sell directly to the public, or does it license independent dealers to do the sales and service? Does the corn flakes company sell directly or depend on supermarkets?

The trend today is away from vertical integration. At many companies, such as Dell and Nike, most of the product is bought from other sources. In other words, these companies coordinate, define, and assemble; they are not integrated.

Banking is one of the most vertically integrated industries anywhere, but that too is changing. These long-term changes, not in any way created by Microsoft, are what really worry people. Microsoft, because of its visibility, symbolizes dis-integration. This trend is advantageous for Microsoft, and for thousands of other companies too.

The total cost structure of banking consists of interest expense, loan losses, and noninterest expense, which the Federal Deposit Insurance Corp. further breaks down into people, occupancy (and equipment), and "other." All these expenses except some portions of "other" occur within the industry, and are therefore vertically integrated.

Portions of "other"-including expenses such as outsourcing the data center, buying software applications, and hiring outside companies to do marketing-occur outside the industry. Such spending has grown more than twice as fast as the internal expenses of the "other" category. It has also more than doubled as a component of total costs, from 4.9% in 1986 to 10.7% last year, and the figure is likely to be 15% to 20% a decade from now.

Costs are not the only indicators of vertical integration. Bank revenue may decline during vertical dis-integration as customers move revenue to nonbanks. For example, the credit card merchant acquiring business was once, circa 1975, predominantly bank work. Revenue was small, and the process at the point of sale was not automated. As electronic draft capture penetrated, volumes and revenue grew dramatically. But much of the business and the revenue moved to specialty nonbank merchant acquirers and processors.

This phenomenon has recurred many times. The volumes of financial transaction processing, financial customer data, and financial reporting have increased by factors of several hundred during the past few decades. Yet the banking industry's noninterest income grew at a compound annual growth rate of only 10.2% between 1986 and 1997. Though this is faster growth than that of interest income, which grew at only 3.3%, the nonbank financial transaction processing companies have grown at about 25% to 35% over the same period.

Why is banking dis-integrating right now? The reason is the power of information technology. Vertical integration lasted as long as it did because record-keeping and processing were paper-based and local. Information couldn't move quickly, it was bulky and awkward to handle, and much of its true meaning was only in people's heads. Note that these factors apply to all information-based businesses, not just banking.

The parts of the industry most likely to dis-integrate over the next five to 10 years are the delivery channels. Banks' 70,000 "stores" cost about $50 billion to operate. All other channels combined (call centers and ATMs) add $5 billion to $10 billion. Securities firms' branches, staffed with full-service brokers, are in the same position. So are 10,000 to 20,000 other financial service delivery locations owned and operated by thrifts, finance companies, mortgage companies, and the like.

Why will delivery channels dis-integrate? Not because some other industry will come in and operate the bricks and mortar. Bank branches won't mimic independent insurance agents. But the evolving electronic information highway distribution system-exemplified by the Internet-will eventually displace most of the cost of today's channels, particularly the bricks and mortar. In its place will come new costs to service the new channels: the costs of participating in the electronic highway.

However, banks will not own the electronic highway the way they do the branches and ATMs. All industries will participate in the information highway, and none of them will own it, except for those in the information- highway-owning business.

This could be very advantageous to banks. Transaction unit costs will ultimately be much lower on the highway, and many of those costs will be borne by others-telecommunications companies, Internet service providers, software companies, PC companies, and, yes, Microsoft.

Some bankers believe they won't "control the customer" anymore. Yet this belief comes with the assumption that bricks and mortar control customers today. It implies that customers cannot distinguish between the channel and the financial service.

This is probably short-sighted. Many other industries have no trouble selling and servicing through channels they don't own. Financial services firms already use indirect channels to raise retirement funds or to sell auto loans. Customers do have to receive value to be sold, but they don't necessarily have to be "controlled."

A second concern is that banks won't get the revenue derived from new financial transaction processing services on the information highway. For example, the bill payment and presentment business will probably be dominated by a few nonbank financial transaction processors. Of course, these same firms will absorb the costs as well. Given banks' dismal revenue growth prospects, why can't banks expand in this direction?

Unfortunately, banks aren't well positioned. A bank, by definition, takes deposits and makes loans. Precisely because of dis-integration, this is no longer the same as doing financial transaction processing. The players in the new areas will probably be smaller, quicker, technology- driven companies, befitting our entrepreneurial culture. Even a big area, like bill presentment, is small relative to the revenue of big financial institutions.

What is to be done about dis-integration? The nonbanks should continue to seek new opportunities in financial transaction processing that result from the trend. Generally speaking, our capitalist system is doing a fine job of motivating firms to develop and exploit new products and services.

For banks and other financial institutions, three strategies are possible. The first is the "ostrich" strategy. Industry dis-integration is a fairly long-term trend, and plenty of profits are possible today. The industry is at historic highs for profitability, so simply ignoring the issue for a while is possible.

The second is the "general contractor" strategy. Instead of just building most things for themselves, banks could begin to assemble pieces of products from elsewhere. This strategy is not restricted to the supply of technology per se, such as outsourcing or buying application packages. Other examples include the sale of "name brand" instead of proprietary mutual funds, sales of loans, and the growing tendency toward indirect channels for loan origination.

The third is the "radical departure" strategy. A new cost structure completely replaces the old one, with no long transition period. Security First Network Bank is an example where the traditional retail channel was eliminated and only the electronic highway used. In this case, the costs and difficulty of switching enough retail customers were high. Security First chose to sell the bank and keep the technology vendor, perhaps preferring the growth side of the dis-integration equation. Firms such as Schwab have a lower entrenched cost base and may be better equipped to adopt this strategy.

For many banks, a move from "ostrich" to "general contractor" may be appropriate. Risks can be controlled; some immediate benefits are possible.

The key change is in attitude. Management needs to recognize the macroeconomic reasons behind dis-integration. It is not a question of banks' lack of vision or the malevolence of selected vendors. In fact, banks' vision should go into thinking about success in a less vertically integrated industry. Banking is not so unique that the experiences of other industries won't be repeated here.

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.