The Top 50 Banks' $53.9 Billion Tech Spend Is Not Enough

This blog is a continuation of last week's blog, which covered the entire population of U.S. financial institutions-16,285, as of Labor Day.Members of the top 50

#1 Bank of America, $2.3 trillion in assets #50 Astoria Federal S&L, $22 billion

The $53.9 billion is a grand total spreadsheet of the 50 IT departments' expense ledgers. It does not include the total cost of 2009's capital expenditures because these are accrued over a three-to-five year period. Thus the $53.9 billion is not total dollars spent, but a combination of spent money (expense) and capital accruals (depreciation and amortization).

For tech vendors who misconstrue all of the $53.9 billion as an opportunity, get ready for a bit of a disappointment. Because the top 50 do a lot of their own IT work, a large portion (42 percent) of the amount is for internal human resources and fringe benefits.

The top 50 spend 72 percent of the total IT expense ($74.5 billion) for all financial institutions.

The IT Core Vendor Community That Views the Top 50 as Prospects

There are 15 bank tech companies that list the top 50 financial institutions as prospects. Seven vendors are already there. Eight would love to be there.

Nine companies are rooted in the U.S., if you allow for a U.S. company that considers Bermuda its headquarters. Three are Indian companies, one of which is owned by a U.S. company. One is a German company. One is a U.K. company. One is a Swiss company.

Any tech vendor with a flamboyant marketing director can correctly claim a top 50 bank as a customer. Even I list 10 of them on my c.v. The top 50 banks buy all sorts of solutions from hundreds of tech vendors. In this context, I'm talking about core solutions. In the top 50, there are 14 banks that rely heavily on a vendor for their core applications. The vendors that provide that support are FIS/Metavante (will become one company on October 1) and Fiserv. Three other vendors provide pieces of core, CSC-Hogan, AFS and Shaw Systems Associates.

A partial core provider is common among large banks. Even though slicing off a piece of the salami defies integration, it is the way large banks evolved. One needs an MRI unit to find integration in a large bank. Everything is interfaced.

Conclusions

The biggest problem that very large banks face is an outdated architecture. Think of architecture as the recent discovery of a crack in the superstructure of the Bay Bridge. One crack could cause a total collapse of the bridge. Functionality is not a problem at large banks. In fact, it's functionality that large banks are reluctant to give up because they'll never get 50 years of custom development from any vendor of a packaged product.

Another problem at large banks is they can't just close down their system for an hour let alone five years while they implement a new system. The Bay Bridge closed for repairs, no doubt resulting in some grumpy commuters. But in a few days, it was crossing as usual. Wells Fargo techies must think of that every time they cross the Bay Bridge.

Remember the last big IT threat the whole world faced? Y2K was a coding problem. Find the date field in a program and expand it to enough bits to show the entire year. Today's big bank problem is architecture. Where do you start? It's like replacing the steel beams in the Empire State Building.

"Big" has its disadvantages. While a community bank can convert to a new system in about six months, and will be operating safely on Conversion Monday, large banks have no safety net or guarantees that all their tests will detect glitches that show up only during production. I believe there are 50 bank CIOs who have never uttered the phrase "Too Big To Fail." "Too Big" haunts them every morning as they look at the stack of trouble reports on their desk.

On the bright side, the current 9.7 percent unemployment rate would drop significantly if a dozen of the top 50 banks were to convert to a new core system. There aren't enough people at Accenture, IBM Global Services, HP, Perot Systems, ACS, CapGemini, CSC, Groupe CGI, Unisys or SAIC that could deliver warm bodies to the conversion projects of a dozen large banks. For a community bank, a call to the local Kelly Services office would take care of the problem.

To bring all this back to the title of this blog, the 50 largest banks are facing a huge increase in cost just to put their legacy systems on a modern architecture. There were few attempts in the past. BankOne, Norwest and EDS tried it. They spent hundreds of millions and several years, and then threw in the towel with nothing to show for their effort. All during those years, the then CEO of BankOne, John McCoy, got more press coverage than a presidential candidate for "leading the pack to a modern platform that would revolutionize banking."

There hasn't been a core conversion at a top 50 bank in decades. And there won't be one for at least another decade since the banks aren't flush with capital these days. The good news is these legacy systems do the job every day without the usual disastrous collapse of a meltdown. I first said it to a group of bankers about 20 years ago: "The easiest IT problem to solve is one that requires money." No one in the room liked that statement. I still believe it today.