The mobilization against terrorism has revived technology investments and added new financial wrinkles to the old military-industrial complex.
By Marianne Sullivan
June 2002
Institutional Investor Magazine
Soon after the September 11 attacks, Oracle Corp. founder and chief executive Lawrence Ellison made an offer he thought the U.S. government couldn’t refuse. Ellison said he would provide his company’s database software , for free , as the basis for a national identification system. He argued that only a gigantic, real-time database, consolidating the many incompatible government agency systems and perhaps accompanied by high-tech citizen ID cards, could plug the security holes that the terrorist infiltration so fatally exposed.
Washington never gave Ellison a formal response , federal procurement procedures work in far more mysterious ways. Nor were Ellison’s motives entirely altruistic. Free software aside, a security system on that scale would require massive expenditures on project management, customization, maintenance and upgrades. Oracle and its many technology and consulting partners would certainly have gotten a piece of that pie.
But Ellison is not alone in thinking that the global response to terrorism could give his company, and indeed the whole information technology sector, a much-needed lift. This different kind of war, entailing both a military buildup and private sector responses , like the financial industry’s actions to combat money laundering , will require billions in new expenditures, many of them for information technology. Input, a market research and consulting firm based in Chantilly, Virginia, expects federal IT spending to climb by 11 percent annually, from $37.1 billion in the current fiscal year to $63.3 billion in fiscal 2007. According to Needham, Massachusetts,based research company TowerGroup, U.S. commercial banks will double their spending on compliance with money-laundering rules this year, to $60 million, and will spend 15 percent more in 2003. Securities firms and other nonbank financial institutions are expected to spend $60 million this year.
The government spending alone could jolt some tech companies out of the depression they have suffered since the Internet bubble burst in early 2000. “Thanks to the federal mandates, we have milestone events, dates certain and huge amounts of money coming in,” sums up Jack Radzikowski, director of emerging markets for Identix, a Los Gatos, California,based supplier of fingerprint verification systems. In a previous job, as head of financial systems at the White House Office of Management and Budget, Radzikowski advocated cutting-edge identification and payment technologies within the government and for consumer disbursements such as welfare and Social Security. Now he wants to help Identix capitalize on the new security demands of airports, military installations and various industries. (The company’s share price recently topped $8, more than double its September low.)
A pair of laws, enacted within two months of the attacks on New York and Washington, is largely responsible for Radzikowski’s confidence. The USA Patriot Act calls for stepped-up money-laundering and financial policing as well as more stringent regulations on the trucking of hazardous materials. The Aviation and Transportation Security Act has, among other things, placed the federal government in charge of airport security. In May President George Bush signed another bill with big technology implications, the Enhanced Border Security and Visa Entry Reform Act.
It is hard to pin down how much tech investment these laws will spawn. Various agencies will have to draw up specifications for the programs, then put them out for bid. But Radzikowski’s notion of “huge amounts of money” is probably not exaggerated. The White House Office of Homeland Security alone has a $38 billion budget, and the newly created Transportation Security Administration, on top of an initial $2.4 billion, may get as much as $8 billion more for this year and next. Additional billions will be tacked on for the departments of Defense and Justice, and much of the money earmarked for antiterrorism and immigration control will go toward technology.
Small figures add up: The trucking, air and maritime transportation and border-crossing provisions in the recent flurry of legislation could require tens of millions of identification cards. If, as expected, these are smart cards with computer chips, they will cost a minimum of $2 to $3 each to produce and issue, a likely boon to companies like Fremont, California,based ActivCard and France’s Gemplus, which have been struggling to develop a mass chip-card market in the U.S.
Introduce smart cards and infrastructure investment is sure to follow: Hundreds of airports and other strategic locations will require machines to read the cards. And companies like Identix that offer fingerprinting, eye scanning and other forms of biometric security are lining up to provide the higher levels of identification and authentication that airports will need to keep track of employees and on-site vendors.
Meanwhile, the Patriot Act has financial institutions scrambling to keep up with new rules for verifying customer identities and bringing suspicious transactions to the attention of law enforcement authorities. Companies like HNC Software, Mantas and Searchspace, which have developed sophisticated pattern recognition technologies that pinpoint a few questionable transactions hidden among millions of legitimate ones, could reap a bonanza.
“We have four to five times more demand for our product than we did pre-9/11,” says James Hayden, head of product management at Fairfax, Virginia,based Mantas, which has recently sold its anti-money-laundering technology to Citigroup and Charles Schwab Corp.
Mantas, which has roots in both the finance and the defense markets, is among the new breed of weapons suppliers in the war on terrorism. The privately held company is peddling technology similar to that developed by its parent, defense contractor SRA International, for the self-regulatory operation of the National Association of Securities Dealers. The NASD has been using the SRA data-mining software, adapted from military uses, since 1996 to flag potential front-running or insider trading.
SRA spun Mantas off a year ago, and in April Mantas raised $17.5 million in venture capital from Wayne, Pennsylvania,based Safeguard Scientifics, which had also invested in SRA. And SRA itself raised $90 million in a May 24 initial public offering.
That’s just the latest in a round of deal making sparked by the war on terrorism. In February Identix moved to broaden its biometric product line by agreeing to pay $261 million in stock for Visionics Corp., a Jersey City, New Jersey,based company that came to prominence after September 11 for its facial recognition technology, which can be used to identify suspected criminals in large crowds, such as at airports. The company recently closed sales to the U.S. Immigration and Naturalization Service and the Belgian federal police. (Visionics’ stock price shot up from $4 in September to $21.12 in December before settling recently at about $11.)
Fair, Isaac & Co., the dominant supplier of automated credit-scoring systems to banks and other consumer lenders, agreed in April to pay $810 million , a 27 percent premium , for San Diego, California,based HNC Software. Fair Isaac primarily wanted to cut out a competitor by consolidating HNC’s analytical software with its own. But the merger is also a homeland security play: HNC’s pattern recognition technology, based on a form of artificial intelligence called neural networking, has both military and commercial applications.
Now, besides applying its software to the money-laundering fight, HNC is collaborating with database vendor Acxiom Corp. and other subcontractors in a Federal Aviation Administration program to analyze passenger lists and airplane routes for potential risk factors. In the same vein, Siebel Systems and Sybase, both competitors of Oracle’s in customer relationship management and database systems, have developed products , Siebel Homeland Security and Sybase PatriotServer , specifically for federal and private sector antiterrorism and anti-money-laundering purposes.
Such systems are designed to facilitate information sharing and analysis, which is exactly what Oracle’s Ellison hoped to do with his free software. It’s just happening one project and agency at a time, with readily available technology that straddles the military and commercial worlds. Allen Jost, HNC’s vice president of business development, explains the connection: “In one case, you are using neural networks to detect what a tank looks like on a radar screen so a pilot isn’t distracted by anything else. In another you are detecting the patterns of fraud. It’s risk management.”
The military-industrial complex that sprang into action after September 11 is not the same megalith created to deal with the cold war’s nuclear threats. Although the U.S. military still relies on such traditional contractors as General Dynamics Corp., Lockheed Martin Corp. and Northrop Grumman Corp. for weapons and other hardware, the 21st- century conflict requires new kinds of expertise. The government needs day-to-day cooperation from financial institutions, the transportation industry and other private sector businesses. And this time the artillery is more likely to come from small entrepreneurial companies.
One thing hasn’t changed: The military can still drive technological innovation. It was, after all, the Defense Ad- vanced Research Projects Administration , Darpa , that planted the seeds in the 1960s for what became the Internet. Going further back, the electromechanical code-breaking machines developed during World War II by British and American scientists gave rise to the mainframe computer industry.
In the old days, however, the government’s massive defense appropriations trickled down to fuel technological development at universities or corporations. HNC got its start that way: Its first customer in 1986 was the Department of Defense, which wanted to harness neural networks , brainlike configurations of computers capable of learning from the information they process. The pattern recognition technology was first used with map readings to detect, for example, enemy tanks amid desert sand dunes.
Defense contracts still account for $3 million to $4 million of HNC’s annual revenues. But the company had $227 million in total sales last year, making it a rousing success in technology transfer , the adaptation of military initiatives for peacetime or civilian applications. HNC’s analytical systems are widely deployed in the retailing industry through Retek, a Minneapolis-based company that it spun off in a 1999 IPO. And HNC’s major customers include financial and health care organizations that use its pattern recognition and predictive software to combat fraud, measure customer profitability and design products and marketing programs.
Now, even as defense spending increases, Washington isn’t just throwing money at speculative research projects. It’s investing in specific technologies to solve specific problems. That funding mixes with private sources, and resulting products flow between the sectors.
That’s the raison d'être of In-Q-Tel, the Central Intelligence Agency’s venture capital arm, which resembles many such firms in Silicon Valley , except that it’s nonprofit. Set up in 1999 to invest in technologies that could benefit national security, In-Q-Tel was wrestling with self-doubt before September 11, recalls chief operating officer Ronald Richard, as technology in general seemed to have turned into the previous century’s obsession. “When you are a start-up yourself and making these decisions for the first time, you kind of wonder if you are really going down the right road,” he says.
The attacks, however, reinforced In-Q-Tel in its mission. Among its holdings are Graviton, a San Diego company making tiny wireless sensors that can gather and transmit data across great distances; Intelliseek, a Cincinnati-based developer of data management and searching tools; and Emeryville, California,based SafeWeb, a provider of enterprise security technologies.
Richard says that In-Q-Tel, which has dual headquarters in Arlington, Virginia, and Menlo Park, California, devotes about a third of its resources to information management and data mining. That serves the interests of the military but also promises to deliver results for banks and other financial institutions, which have been struggling for years with database management problems similar to those that Oracle’s Ellison wants to solve for government.
Explains In-Q-Tel’s Richard: “Most of what we do is enterprisewide software solutions for the CIA. And any large entity, like a Citibank, also needs these as they are dealing in different countries; they need things automatically translated, collated. They would need to do deep searches to check out the financial status of a company they wanted to lend money to. All of these things are applicable, not only to the financial services industry, but to insurance companies or any large enterprise.”
The similarities between government and financial data management help explain why HNC Software, when it was mapping plans to diversify out of the defense sector, went most aggressively after financial institutions. Business development director Jost says that when he joined HNC in 1990 as it began that push, banking was decades ahead of other sectors in its understanding of how pattern recognition and statistical analysis could benefit business operations.
Indeed, deep searches, data mining and analytics are now all the rage. John Frank, a Massachusetts Institute of Technology student, obtained Darpa funding last year for a data-mining system that he developed as part of his graduate studies in physics. He is hoping to sell his technology to financial institutions as well (see story below).
Another financial markets example is NetExchange, a 1994 spin-off from California Institute of Technology, where faculty members designed a computer system to allocate research money among the school’s scientists at the Jet Propulsion Laboratory. The underlying technology, called computational combinatorial deal making, in effect created a marketplace for competing ideas. It has since been applied in the bond market through State Street Corp.'s BondConnect electronic trading platform and in the trading of environmental emissions credits.
San Diego,based NetExchange is now working under a Darpa grant to turn its technology into a foreign-policy analysis tool: Experts, including government analysts and academics, would place bets on outcomes or scenarios , buying “shares” in the probability of war between two countries, for instance.
“People will come in with different snippets of information and in the end discover prices that are better signals of what is going to happen than any they had individually,” predicts NetExchange president Charles Polk. He compares such risk management tools to the orange juice futures market: “The price of a three-month contract is a better predictor of the weather in Florida than the National Weather Service.”
Technology transfers from military to financial, and vice versa, are not limited to neural processing, security and market making. Nor are they all homegrown. Russian rocket science forms the core of Egar Technology, a fast-growing provider of derivatives-related technologies to Wall Street (see story below).
Also with an eye on securities markets, U.S. defense contractor Anteon International Corp. is embedding video compression technology, originally developed for signal processing on submarines, in handheld computing devices for stock traders. Fairfax, Virginia,based Anteon, which raised $270 million in a March 21 IPO, has spun this business off as the Pocketmultimedia division. The unit is currently working on two Darpa-funded projects , “robo bees” that can swarm into enemy territory and transmit observations back to battlefield commanders, and miniaturized remote-control surveillance vehicles. Pocketmultimedia handsets, which can handle videoconferencing and other large-bandwidth data feeds, have been tested by New York Stock Exchange floor traders.
“The government realizes that custom-built systems are costly to maintain and knows that if it wants fast innovation, it needs to look at commercial off-the-shelf technologies,” says Robert Manchise, Anteon’s chief scientist.
Two off-the-shelf technologies , smart cards and biometric security systems , may be about to get their long-awaited big break. Manchise says that he expects smart cards, already incorporated in high-end, military-grade security systems, to be repackaged for broader use. The cards originally gained popularity as stored-
value payment devices for public telephones in Europe but have been slow to win acceptance in the U.S.; American Express Co.'s Blue card is a notable exception. Smart cards are now seen as relatively cheap storage media for ID and authentication data ranging from digitized photographs to fingerprints to encryption codes.
The Defense Department began a two-year project last fall to issue 4 million multiapplication smart cards , for authentication, payments, battlefield logistics and other purposes , to military and civilian personnel. The Immigration and Naturalization Service’s technology upgrade could require more than 20 million smart cards to keep track of people with visas. The Transportation Security Administration’s campaign to register workers and to secure facilities at airports, seaports and other locations could call for as many as 15 million cards.
In addition, the 50 states may adopt smart cards for the nation’s 200 million drivers’ licenses, which are now all too easy to counterfeit. A bill currently before Congress would mandate smart cards, and the American Association of Motor Vehicle Administrators, which represents the state authorities, at the very least wants to raise standards for verifying applicants’ identities as well as licenses’ validity. As the state officials get into the nitty-gritty of how to upgrade, they intend to call in people who are experienced with ID and authentication issues. Says Linda Lewis, president and CEO of the Arlington, Virginia,based AAMVA, “Financial institutions will definitely be at the table.”
There is more to tracking money launderers and terrorists than just monitoring transactions, according to Bernard Azoulai, a managing principal at New York,based financial services consulting firm Capco. “You have to see if there is a link beyond your customer to an outside company that could be the money launderer,” he says. “One of the problems is that money launderers are much smarter than they used to be.”
For all the technology that is out there to profile customers and track banking patterns, systems capable of finding the links between suspect transactions and terrorists have yet to be perfected. Until they are, the new war is likely to require much more technology investment , perhaps even in Ellison’s Oracle.
Laboratory to military to finance: An MIT case study
How does a Ph.D. candidate in physics, immersed in the arcana of condensed matter theory, become a defense contractor? And how does he make the leap from there into marketing technology to the financial industry?
Just ask John Frank, for that is the improbable life this 25-year-old has been leading in the 14 months since the U.S. Department of Defense’s Advanced Research Projects Agency got wind of his studies at the Massachusetts Institute of Technology.
He started out innocently enough, trying to find clusters of trees whose emissions of moisture could seed their own rainfall. To aid his search, Frank wrote software to filter through vast amounts of geographical data and pinpoint promising clusters. He had an inkling that his application of geographical data mining , akin to the systems that financial institutions and other corporations use to analyze customer databases , might have some commercial appeal. So in January 2000 he formed a company, MetaCarta, to serve as his marketing vehicle.
The start-up flew beneath all radar screens until April 2001, when, on a recommendation from Frank’s faculty adviser, a Darpa project officer invited Frank down to Washington, D.C. The following week Frank demonstrated his system there using a wireless handheld device, and within two months he had a $500,000 grant from the agency to develop a prototype that the military might apply in its own data and intelligence gathering.
“If you take a step back, the original tree-cluster probings at MIT are a lot like what we are doing right now,” Frank notes.
That work, by Frank and a ten-person team, involves creating programs that find geographical cues and correlations within unstructured text, such as e-mail messages or news articles. Text mining is in hot demand by everyone from Central Intelligence Agency country analysts poring through reams of published material, to mortgage bankers extracting pertinent data from loan applications. The same technology that military strategists might use to locate land mines or enemy antiaircraft installations can help financial companies track the assets in clients’ investment portfolios and manage risks around the world.
The MetaCarta technology finds geographical points of reference, just as someone reading an e-mail might deduce where it came from, Frank explains."However, teaching software to pick up on those references and build the same associations is hard,” he adds.
With sufficient data inputs, his system can plot a set of targets at specific latitude and longitude positions on a map, within a margin of error or confidence level that a user specifies. Credit card companies do similar things with consumer demographics and creditworthiness data, which they use to devise product offers at various interest rates and risk tolerances.
Since the September 11 terrorist attacks, the government and financial sector alike have stepped up their demands for sophisticated analytical tools , whether for rooting out enemies on battlefields or for complying with anti,money-laundering regulations that require banks to keep tabs on the sources and flows of customer funds.
Frank sees his technology filling both sets of needs. He hasn’t deployed any products at financial institutions yet, but he has held serious talks with at least one big Wall Street bank that wants to do a better job analyzing and managing the geographical risks in its portfolios.
Frank is aiming to complete a first round of venture capital financing in the next 12 months. He sees In-Q-Tel, the CIA’s venture capital arm, as a logical and likely participant. Frank has also been working through In-Q-Tel to get CIA officials to consider MetaCarta software. And his Defense Department connections have opened doors into operational areas of the armed forces.
“It is a way for a company like mine to better understand what kinds of technology they might be looking for,” says Frank. “And then there is the flip side: Since they don’t always know what technology is available, it is for us to show them what is possible.” , M.S.
From Russia with code
Egar Technology bills itself as one of the fastest-growing software and services providers to Wall Street. The company has supplied trading and risk management systems to Bear, Stearns & Co., CIBC World Markets and Société Générale’s Fimat USA subsidiary, among others. Through an affiliate, IVolatility.com, it delivers market data and analytical tools to many of the biggest derivatives trading desks.
But Egar’s small Midtown New York sales office is just a front. All but 12 of its 122 employees work at the company’s systems development and engineering hub , in Moscow.
Though India has gotten much notice as a hotbed of offshore software development and support, Egar deemed Russia a better fit.
Moscow is teeming with scientists and engineers who were left underemployed by the passing of the cold war, says Egar CEO Ravi Jain. One of them, Egar head financial engineer Mark Ioffe, spent 25 years working for the Moscow Research Institute, a division of the Russian Department of Aviation, where he developed missile guidance systems. He now applies his mathematical expertise toward developing models and formulas for derivatives transactions.
Ioffe and other Egar recruits, most holding advanced degrees, nominally work for a Russian-domiciled subsidiary, but they are busy transferring technology to the West. “I couldn’t compete for people in India. We had a much better chance in Russia,” says Jain, 40, a former head derivatives trader at Republic National Bank who launched the technology company in January 2000.
One of India’s attractions had been its English-speaking population, but Jain, himself of Indian descent, says the Russian talent pool is superior , and speaks English. But it also happens that Jain’s co-founder and chief technology officer, Gena Ioffe (Mark’s nephew), is Russian. The 42-year-old Ioffe, who developed cutting-edge derivatives technology in the 1990s at transaction systems vendor FNX and at a company he started, Delta Analytics, earned his BS degree in computer science and an MS in applied mathematics from Moscow State University.
“Russia has gotten very hot for outsourcing,” Jain notes. U.S.-based companies like Boeing Co., Intel Corp. and Microsoft Corp. are there, but Jain says: “We’re unique as a capital markets company. For complex applications like derivatives, Russia is a perfect skill set , people all over the place with master’s degrees in math, physics and engineering.”
Jain says that Egar’s Moscow office could be an eventual base for expansion in Europe, but for now the company remains focused on the U.S., targeting small and midsize financial institutions as well as the specialized trading and risk management departments of large banks and corporate treasuries.
Egar sells itself as neither a pure vendor of prepackaged systems nor a pure consulting firm that integrates others’ systems. “We can do 90 percent our products and 10 percent custom development, or the reverse, or mix and match anywhere in between,” says Jain. He is confident that the firm can maintain that flexibility , as long as the Russians keep churning out the code. , Jeffrey Kutler
