Opening windows with the internet
From the archive: in June 1996, with some prescience, then Euromoney reporter Felix Salmon investigated how far the net, or "internet protocols" would shake up banking.
The internet’s free technology raises questions about all the old ways of doing things. It makes communication easier than ever – both inside and outside a company – and it revolutionizes information storage and access. But who will be first against the wall when the changes begin to bite? By Felix Salmon.
Joe Banker is feeling old – and he’s only 24. He’s just had a brand-new Sun workstation installed next to his Windows 95 PC and is unsure about what it all means. It’s got one of those clever optical mouse things attached to it. It boasts three buttons. Bill from IT was ever so enthusiastic about it in the pub last night. “Functionality,” he explained, “is all about what your machine is capable of doing. The analysts now create beautiful reports in Word 7, embedding Excel 6 charts. The rocket scientists can run extremely powerful options-pricing algorithms in half the time it would have taken them last week. And if you press shift-control-A and the middle mouse button, an instruction is sent to the coffee machine to have a fresh cup ready for you in four minutes’ time!”
Joe’s stomach begins to tighten whenever anybody comes along to install a new piece of hardware or software. He knows there’s a wonderful internal computer system which makes his life a lot easier, but he just can’t understand all the different commands, applications and error messages which appear whenever he tries to use it. Most of the time, the only thing on the screen of his ludicrously over-specified workstation is the CNN home page on the World Wide Web. That, he can understand. It won’t plot a yield curve, but it gives him what he wants.
But soon, Joe is going to come across a technological innovation he’ll actually welcome – an internal internet, or intranet, which provides a single interface for software, information, communications, etc. If 1995 was the year that banks went into the internet, then 1996 is the year that the internet goes into the banks.
At Morgan Stanley, every employee’s computer is already fitted with a Web browser. Logging-on brings up an internal home page containing internal news that is updated daily. All in-house equity research is accessible from internal servers, which drastically reduces paper and printing costs, and saves vast amounts of filing cabinet space, all over the world. Some facilities are marginally used – for example, a search engine which identifies members of staff on Morgan Stanley’s office floor plan. Others, such as a completely up-to-date internal telephone and e-mail directory, are used all the time.
There is a comprehensive and well-indexed list of Web sites relevant to employees in every imaginable part of the bank, which are available with one click of a mouse. There are discussion groups on a large number of subjects – related and not-so-related to work – to which employees all over the world contribute via e-mail. And any information that Morgan Stanley wishes to make available to clients or to the general public can easily be placed on its Web site, tagged so that all or only some people can have access to it.
The people running the system can easily tell which services are proving popular and which aren’t. Mark Kennedy, a former Morgan Stanley technology vice-president, has called this its “most overlooked benefit”. This information, he writes in a research briefing, “is invaluable. It separates the winners from the losers in the information world. Once something can be measured, there is almost no end to the ways the measurements can be used.”
All of this marks a profound shift from the over-specification which is the bane of Joe Banker’s life to, if anything, under-specification. His old desktop computer, in theory, enabled him to do about 905% of what he wanted to do but the difficulty of charting a course through the maze of options tended to restrict him to far less than 100%. Now with little more than a point-and-click Web browser, he can now do about 95%. Soon, this figure will get closer to 99%, thanks to the internet’s hottest new piece of software, Java.
Java, the internet’s own computer language, is a key element of its accessibility and appeal to users. Whenever there’s something Joe wants to do, the IT people can write a simple Java application that allows him to do it.
The openness of internet protocols means that it doesn’t matter if the employee is working on a Unix workstation or a PC – or even an Apple laptop in a hotel in Tajikistan. An in-house application written in Java can be used by any computer accessing the network. Until very recently, each application would need to be written once for each system.
Moreover, there are many reasons, according to Morgan Stanley’s Kennedy, “to think that applications written in Java will be better than the standard alternatives”. For one thing, they’re invisible: the user still has only the point-and-click Web browser interface to learn. For another, applications (and their updates and patches) don’t need to be installed on every machine in the company: they just sit on a single Web repository until needed.
This is the major benefit of the internet for finance professionals: ironically enough, it actually reduces information overload. It marks a move from a “just-in-case” to a “just-in-time” approach to information storage. No longer is it necessary for everybody who might need a piece of information to have a copy of it. Instead of massive duplication of information in both paper and electronic form, internet servers can provide users with the most up-to-date version of any document they want, when they want it, whatever machine they are working on, wherever they are in the world.
Bandwidth is the thing
The standard “look and feel” of a Web browser and its ability to work on any old machine is a big step on from standard client/server systems. The user no longer needs to understand a multitude of applications and certainly doesn’t need to have a copy of every application he’s likely to use installed on his machine. (In a recent survey, 56% of managers said that this sort of ease of use was the single most critical factor for a corporate information system.) Indeed, the user doesn’t even need a PC at all: a standard $500 “network computer” of the sort announced last month by Oracle, equipped with a very basic chip and internet connection, should suffice. Information professionals – which include bankers – no longer aspire to having faster chips than the next man: the important thing today is bigger bandwidth.
Furthermore, an internet or intranet (the new generation of internal network) system is much more versatile than an old-style centralized information database and interface. Instead of relying on programmers to create data entry and reporting programs, information can be collated, published and managed at source. In short, control of information no longer resides in one place.
Large, centralized information providers such as Reuters, Bloomberg and Knight-Ridder are basically editors and compilers of information they are paid both to receive and to distribute. The internet shows one way of cutting out the middleman altogether and “disintermediation” is becoming as important a word in IT as it is in banking.
The number of “hits” (downloaded files) made by Morgan Stanley’s 9,500 employees on its intranet has increased from nothing three years ago to 1.6 million per week now, with no sign of slowing down. These numbers dwarf the 10,000 to 20,000 hits per day received by Morgan Stanley’s Web site – the information available to the rest of us – impressive though that number is. What is really happening is that the growth of intranets in banks is providing the demand for which external sites are forming part of the supply. As the former grows so will the latter (including among them Euromoney’s own which goes on line this month) and in the process large parts of the information industry will be restructured.
The Morgan Stanley Web site is popular for two reasons: it provides high-quality information unavailable elsewhere and it updates it regularly (daily, in most instances). Its site includes the Global Economic Forum, with analyses of global economic trends, as well as equity analysis for clients.
But although there are one or two Web sites of interest to many financial professionals, the World Wide Web is at heart the apotheosis of narrowcasting: the information tends to be so fine-grained that only a small number of people access any particular page.
The internet started life at the US Department of Defense’s Advanced Research Projects Agency. Most internet backbones – the fibre-optic networks that the internet runs over – were built by western governments. The World Wide Web was developed at the Conseil Européen pour la Recherche Nucléaire (CERN). Today’s most popular Web browsers are more or less the same as those created at the University of Illinois’s National Center for Supercomputing Applications.
Users of the internet are therefore leveraging large amounts of public money. But increasingly nowadays, the information access from the internet is no longer free. The death knell of the publicly-funded internet was sounded in April 1995, when America’s National Science Foundation announced that it would no longer fund the NSFNet T1 backbone. It passed the responsibility for infrastructure maintenance to six commercial companies – PSINet, UUNet, ANS/AOL, Sprint, MCI and AGIS-Net99. These companies now control the internet’s fibre-optic motorways.
The first users to notice the introduction of charges were those wanting to register a domain name (the identifier which comes after the “@” in an e-mail address). But then regular Web surfers began to notice a very large increase in corporate advertising on popular sites and an explosion in the number of sites requiring registration and/or a subscription fee.
Two years ago it would have been unthinkable for the Banking and Finance it Informer newsletter, distributed over the internet, to charge its current annual subscription of $300 for the service. Already many popular sites, such as the Financial Times and HotWired, require registration before they can be accessed. Registration usually adds no value for the user (although “Daily Me”-style personalized pages are becoming more common), but gives the provider valuable demographic information.
Salomon Brothers now charges for information it makes available on the World Wide Web, but it is unlikely that the internet will be a significant source of revenue for investment banks in the foreseeable future. “There’s not a great number of opportunities to make money, but there is a great number of opportunities to serve clients better,” says Philip Harrison, head of Morgan Stanley’s European internet activities. Clients who use the bank’s Web site are much more likely to give that bank valuable mandates.
Banks were slow to pick up on the availability of the free technology, largely because they tend not to believe in the “free lunch”. Another reason was security: open networks are always less safe than closed ones. The computing industry was happy to co-operate in this: although suppliers paid lip service to open networks, they wanted everybody to use their protocol. It took disinterested academics to create truly open protocols for use on the internet.
Recently, however, banks have been embracing internet technology with all the zeal of the newly-converted. In a recent survey for Mercury Communications, 92% of financial institutions in London and 88% in New York said they were either using or planning to use the internet. For some, this involves little more than a dial-up account with a retail internet service provider. For others, it forms the backbone of their entire corporate it system. And when City Research Associates polled City financial institutions last month, 62% said, unprompted, that the internet would play a very significant role in their future.
Banks, however, are not entirely happy. Some have yet to be convinced of the solidity of internet technology. “We don’t find the software reliable,” says Mukarram Sattar, head of operations, technology and software at Citibank in London.
Other banks use Lotus Notes for internal communication and make it deliberately difficult for employees to send external e-mails. (Lotus Notes can do more things than e-mail or most intranets, but just as with the PC, this is not always a good thing.) Much internal communication in a bank is highly confidential and e-mail is the perfect communication tool. Bankers spend a lot of time on the telephone, but ringing up a colleague is not always an easy procedure: there can be very few financial professionals without a deep-seated loathing for voice mail. An e-mail, by contrast, appears as soon as it is sent, can include detailed data and lasts in unambiguous written form for as long as it is needed. What’s more, it can be sent to a number of people simultaneously.
And this is where the problem lies. A desktop e-mail application can send highly confidential data from Fred@jpMorgan to Joe@Salomons just as easily as it can to Fred’s colleague John in the office next-door. Faxes have been arriving where they weren’t intended to go for many years now, and many cyber-denizens have received e-mails not meant for them.
There is encryption technology, of course. Morgan Stanley has an encryption program for sending confidential information outside the company that is so powerful it needed to obtain a munitions licence from the US government before being allowed to use it. But widespread use of such technology within a company is enormously wasteful of bandwidth and is rarely used.
Certainly, “you’re not going to route internal communications via the internet until capacity and encryption techniques improve”, as Morgan Stanley’s Harrison says. In other words, hackers aren’t going to be able to intercept your internal mail en route. But Harrison is less convincing when he says that external people won’t receive confidential information by mistake because their e-mail addresses require a domain name. (For example, Jack at Morgan Stanley can send an e-mail to Gary at Morgan Stanley just by typing “Gary”; to send it to Fred, he needs to type in “Fred@JPMorgan.com”.) The fact is that no e-mail program requires the full e-mail address of the recipient to be typed in each time. And by the time Jack notices that one of his recipients ought not to have been on the list, it’s too late.
Another common worry about providing all employees with internet access is that they will waste valuable time aimlessly surfing. It’s certainly easy, even for a surfer with the best of intentions, to get sidetracked or “caught in the Web”.
Anecdotal evidence suggests that the amount of wasted company time is quite high for the first couple of weeks that an employee is connected, but rapidly levels off thereafter. Still, says Harrison, “we need to be pro-active about policing it”. Morgan Stanley employees, for instance, cannot visit the most popular Web site on the internet: that of Playboy magazine. Pro-active maybe, but not Draconian. After Morgan Stanley blocked access to the ESPN television sports network’s site after noticing particularly high usage of it, it received a barrage of complaints from salespeople and traders. Apparently, many would use the sports results as a reason to phone clients: “Haven’t the Bulls done great? Oh, and while I’m here, do you think I could interest you in a Kingdom of Sweden 10-year at eight under?” Access was soon restored.
Paul Lindner, one of the architects of Gopherspace – the briefly-popular precursor to the World Wide Web – has said that “distributed computing is like driving a wagon pulled by a thousand chickens”. The act of harnessing these chickens is not an easy one, and banks are now competing with the likes of Netscape, Microsoft and IBM for the computer geniuses who can do it well. They don’t come cheap. (Conversely, technology companies are hiring bankers in an attempt to get lucrative hardware and software contracts. In the City Research Associates poll, 79% of financial institutions spending more than £10 million per year on IT said that external IT suppliers were not sensitive to their business requirements.)
The rise of the internet marks the democratization of IT. Previously, it took mind-boggling amounts of money to exploit state of the art technology fully – the largest investment banks each spend upwards of $500 million annually on IT. Now, anyone with a PC can join this information elite.
“In the last four years, the investor base for Brady bonds has broadened significantly from one that was almost completely dominated by professional traders to one with a significant proportion of retail investors,” says the introduction to one new site on the Web. “Until BradyNet started publishing Brady bond information, technology had not kept pace with that very important demographic sectoral change and virtually no new information channel had opened to improve market transparency for those investors.” In this respect BradyNet is typical of many new sites.
Just as banks are cutting out the middlemen in many areas, their role as middlemen is also changing. Banks have made money by providing bond and equity prices and analysis to small and medium-sized investors. And they have made even more by providing, say, risk analysis for the same clients.
Now small technology companies, such as Palo Alto-based Integral, are offering companies an alternative way of managing derivatives, and getting advice and pricing. It makes use of JP Morgan’s RiskMetrics data which is freely available on the internet and then offers, according to the company, “objective accurate pricing and Value-At-Risk (VAR) exposure analysis for virtually any type of derivative or hedge instrument. Portfolios can be analyzed for VAR, current mark-to-market and other customized reports. Reports can be run in real time or processed for overnight delivery anywhere in the world via e-mail or fax.”
“We are targeting corporates who will spend $30,000 to $50,000 a year on RiskNet,” says Raj Patel, Integral’s managing director in London. “Typically, a bank rings up a corporate and says it should do a swap. The bank then transacts that swap with that corporate. We can provide a second opinion.”
Patel would like to keep banks happy too: for about $3 million a year, he’ll let them outsource their risk management advice operations and provide software for their corporate clients, with built-in profit margins, using whatever model they’d like.
Integral can take heart from the City Research poll: 79% of financial institutions spending more than £10 million per year on IT said that they would increase their spending on external packages over the next three to five years.
On balance, the internet must be a good thing for banks. It can revolutionize both their internal and external communications. It makes technological integration after a merger a much less painful process. It provides opportunities for outsourcing much expensive technological legwork to companies such as Integral or The Bankers Network. And it brings them closer to clients. But whatever privileged position banks had in the old information culture is eroding fast.
The Mercury Communications poll came up with another very interesting result. When asked what the most important element of a business was today, financial professionals in Tokyo, Paris and Frankfurt voted two-to-one in favour of human resources over technology. In London, however, technology won by 41% against 31%, and in New York the ratio was 31% to 11% in favour of the computer. It would seem that, in the world’s biggest markets, a bad employee with good tools can do more than a good employee with bad tools. And the internet is now the world’s most powerful enabling tool.
The internet is here to stay and will grow in importance at its present exponential rate for the foreseeable future. Already, Wells Fargo offers full retail banking services over the internet. It is only a matter of time until share trading joins in too. Eurobond traders are probably safe: big deals need the personal touch. There will also be an explosion in the number of bankers telecommuting: there is little reason for dealers, say, to work in expensive city centre offices.
Telecommunication companies ought to be worried: the technology already exists for telephone calls to be made at zero marginal cost over the internet; with the next jump in bandwidth, this could become commonplace. After that comes speech recognition and television-quality real-time video and voice applications. And after that, we might finally see the first virtual banks.