An unenviable position

When James Ziglar envisions the future of the agency he was appointed to reform, he imagines an Immigration and Naturalization Service equipped with computer systems able to closely track the 7 million foreign visa holders who arrive in the United States each year.

Today, without an effective entry/exit system, millions of visitors stay on undetected after their visas expire.

Another automated system INS Commissioner Ziglar foresees would monitor the 500,000 foreign students granted visas each year to study in the United States. The student tracking system would check that visiting students comply with a host of laws that govern their stay, from what they can study to how many hours they can work. The system is also intended to ensure that student visa holders are, in fact, students.

With the limited information it maintains on student visa holders today, INS was unaware that at least three of the Sept. 11 hijackers were among them.

In Ziglar's vision, visas would include biometric identifiers, could be read by machines and would be highly resistant to counterfeiting.

And INS computer systems would be linked to those of other agencies so that INS inspectors, State Department consulates and others could instantly share information, making it harder for criminals and terrorists to enter the United States.

Technology could be "a huge force multiplier" helping INS accomplish its mission, Ziglar contends — if the agency could use it effectively.

But INS has a long history of technological ineptitude. In an era when other government agencies embrace Web-based services, INS continues to struggle with paper documents. Amid the increasing need for real-time data, INS' multitude of old computer systems cannot communicate with one another, let alone with the outside world.

All of which leaves Ziglar in a difficult position. It's clear what needs to be done, but given his agency's poor track record, it's not clear if he can do it.

Sizing Up the Challenge

INS is "big on information and small on technology," Ziglar told the House Judiciary Committee in March.

He was called before irate committee members to explain how student visa approvals were mailed to two of the Sept. 11 hijackers six months after they were killed in the terrorist attacks.

Ziglar described a Byzantine process in which INS employees collect paper visa documents in boxes, and when the boxes are full, ship them to a contractor, where information from the paper forms is typed into databases. The process can take more than six months.

Ziglar, who became INS commissioner in August 2001, said he inherited "antiquated technology systems and overly bureaucratic processes."

"There's no debate as to whether the INS must be reformed," Ziglar told the Judiciary Committee in another appearance April 9.

A closer look at the critical systems that Ziglar and others say INS needs reveals the size of the challenge.

INS' efforts to create an automated entry/exit system for tracking foreign visitors consumed five years and more than $31 million, but produced a system that worked at just four airports with the voluntary cooperation of only two airlines. "INS has not properly managed the project," the Justice Department's inspector general reported.

As for monitoring foreign students, INS today essentially cannot.

"Today, the INS maintains limited records on foreign students," Ziglar recently told lawmakers. The data resides "on old technology platforms that are insufficient for today's need for rapid access," he said.

Nor has INS demonstrated competence with biometric identification systems. Since 1994, the agency has used a fingerprint and photograph database known as IDENT along the border with Mexico. Its job is to store fingerprints and photographs of illegal and criminal aliens so authorities can be alerted if they try to enter the United States.

But an investigation by the Justice Department's inspector general found that fewer than one out of three aliens apprehended on the U.S./Mexican border were being fingerprinted, photographed and entered into IDENT.

The system itself works well, according to a government investigator. But INS personnel, overwhelmed by a tide of illegal migrants crossing the Mexican border, often did not have time to take fingerprints and photos and enter them into IDENT.

The inspector general found that training for IDENT "was ineffective or nonexistent" and that there were "virtually no controls in place to ensure the quality of data being entered into" the system.

On June 1, 1999, the lack of data in IDENT prompted the Border Patrol to release a Mexican national caught trying to cross the border into New Mexico. Nothing in IDENT alerted officers to the fact that Raphael Resendez-Ramirez was wanted as "the railway killer." He slipped back into the United States and murdered four more people before surrendering July 13, 1999.

"IDENT has gotten a bad rap," said Scott Hastings, INS' associate commissioner for information resource management. IDENT and an associated case management system, ENFORCE, have led INS to apprehend "some real bad guys. It doesn't always hit the papers," he said.

In another effort in 1998 at biometric identification, INS began issuing "laser visa" cards to replace paper border-crossing cards used primarily by Mexicans who enter the United States frequently.

More than 5 million of the plastic credit-card size "visas" have been issued. They contain identification data, including fingerprints, that can be read by a machine. But so far, INS has installed few of the needed readers.

"The smart cards have been effectively rendered dumb cards," Sen. Dianne Feinstein (D-Calif.) said. Feinstein is pushing for passage of legislation to require that all visas be machine readable, tamper resistant and contain biometric identifiers.

The problem was cost, an INS official said. INS did not have money for the readers. Now, the laser visa program is expected to be subsumed by a new entry/exit system.

A Catalog of Problems

Both Congress and President Bush have ordered INS to develop "a fully automated entry and exit control system" to keep track of legal foreign visitors. The system is to be installed at airports and seaports by the end of 2003 and at all entry points by the end of 2005.

It won't be an easy task.

"INS has had long-standing difficulty developing and fielding information systems to support its program operations," said Richard Stana, director of justice issues at the General Accounting Office.

In a steady stream of audit reports spanning more than a decade, GAO has compiled a catalog of INS system deficiencies. There is a common thread among them: INS lacks an overall technology architecture, said Randy Hite, chief of GAO's information technology architecture and systems issues department.

INS' approach to technology has been that "where a problem came up, they developed a solution for that problem. People didn't think of how to optimize for the whole. They thought of how to optimize for my part," Hite said.

As a result, INS developed most of its automated systems independently from one another. Many of the systems therefore cannot communicate to share information, so data must be entered into each separately. That means duplication of work, delays and mistakes.

INS lacks an enterprise architecture, Hite said in a report to Congress in October 2001. The agency spends about $300 million a year on IT but without a plan that would make systems work together, he said.

Hastings concedes that INS has serious problems with "interoperability, knitting together" its disparate IT systems.

"But I'm not sure we deserve failing marks on systems," he said. When investigating the Sept. 11 terrorists, "we were able to recover information on the activities of most of them" by searching through system after system.

"We knew when Mohamed Atta exited and entered" the country and when he applied for changes in visa status. But timeliness was a problem. The information could not be found easily or quickly. But the INS systems date to the 1980s. "When they were developed, that's the way it was done," Hastings said.

INS efforts now are aimed at making information retrievable through many systems, he said.

Since last fall, the agency has made modest progress toward developing a technology architecture, Hite said.

'Management Deficiencies'

But it will take more than an IT architecture to solve INS' technology troubles, according to Justice Department Inspector General Glenn Fine. Repeatedly, the Office of the Inspector General "has found serious process and management deficiencies in the INS," he said.

All too often, projects are inexplicably delayed, costs "spiral upward with no justification" and projects proceed with no assurance that they will ever perform well or meet agency needs, Fine said.

Much of the poor management occurs because many "key managers aren't systems people," said a government official who has scrutinized the agency's IT problems in detail. Technology projects are often managed by career agency employees who have only a vague acquaintance with technology and who view the assignments as a way to "get their ticket punched" so they can qualify for more senior-level management jobs, he said.

One example is the Border Patrol's Remote Video Surveillance program, he said. Cameras mounted on towers enable the seriously understaffed Border Patrol to monitor large sections of the U.S. border. "The Border Patrol is good at apprehending people, but it's not good at systems," the official said. "They're not getting the right kind of cameras. There are much more advanced systems out there."

Ziglar said he plans to solve such technology management problems by hiring a chief information officer for INS.

But Hastings warns that technology without adequate personnel to use it "isn't going to get us too far." INS has about 4,775 inspectors to check 500 million people who cross U.S. borders and enter U.S. airports and seaports each year. And the agency has 2,000 investigators to deal with people who have entered the United States illegally, overstayed their visas, committed crimes or violated the terms of their visas.

"When you look at the range of the mission we have to carry out," INS' staff is stretched thin, Hastings said. The northern border alone is 4,000 miles, he said. At one spot on the southern border, San Ysidro, Calif., 50 million people cross each year.

"Clearly, technology has to be utilized to control migration across our borders," Hastings said. "But someone has to be there to utilize the information" provided by technology.

Pending legislation would enable INS to hire more inspectors and investigators. The agency also hopes to rely more on states and localities for law enforcement help.

Congressional Review

Although INS has been an inviting target for congressional critics for years, it became even more so after the Sept. 11 terrorist attacks. But many who are familiar with INS say that Congress is responsible for some of the agency's deficiencies.

For example, Congress voted in 1996 to require INS to develop entry/exit and foreign student-tracking systems, but twice since then has voted to delay the systems. Under pressure from groups representing colleges and universities, employers and business interests, lawmakers pushed the deadline back to January 2003.

Associations representing chambers of commerce and other commercial interests worried that systems designed to tighten control over the nation's borders might also slow commercial traffic coming from Canada and Mexico.

Meanwhile, employer organizations opposed the development of a computerized verification system that could determine the eligibility of foreigners to work in the United States. "They were worried that it would actually function," said Mark Krikorian, executive director of the Center for Immigration Studies. A functioning verification system would certainly have shrunk the pool of available employees during the worker shortages of the late 1990s.

Colleges and universities, too, opposed the extra work and intrusiveness of reporting information on foreign students to INS.

Congress backed down, and INS continued to operate with ineffective systems, Krikorian said.

Until the terrorist attacks, INS was under fire mainly for its inability to promptly process visa requests, citizenship applications and the like. At its worst, in the late 1990s, it took INS 28 months to process a typical naturalization application and as long as two years to grant permanent resident status. The backlog of unprocessed applications climbed into the millions.

Under pressure from Congress to speed things up, INS concentrated on completing paperwork and was less stringent about scrutinizing applicants, said Jessica Vaughn, a former State Department visa officer stationed in Spain. In some instances, approval or denial of applications was left up to data entry clerks, she said.

Now the emphasis at INS has shifted from faster processing of paperwork to greater homeland security. And again Congress is intervening.

IT on the Fast Track

Legislation pending in the Senate mandates greater use of technology.

Sponsored by Sens. Feinstein, Ted Kennedy (D-Mass.), Jon Kyl (R-Ariz.) and Sam Brownback (R-Kan.), the bill requires systems to be built for sharing information among law enforcement agencies and for electronic visa files to be shared by the State Department and INS.

The bill, which was debated last week on the Senate floor, also spells out technology standards for electronic visa systems to be used at border crossings, for creating an entry/exit system and for ensuring the interoperability of databases involved in determining the admissibility of aliens.

Under the legislation, INS and State would be required to issue machine-readable travel documents with biometric identifiers. And countries for which visas are not required would have to issue passports that are machine- readable and include biometric identifiers.

"This legislation would help transform the Immigration and Naturalization Service from a paper-driven bureaucracy," Feinstein said. It would give INS officials access to "vital intelligence information in real time before they issue visas and permit entry into the United States."

"I've been with the INS for 18 years, and I don't know that there's another federal agency that has had as much legislation passed" pushing it in one direction or another, Hastings said.

"It's naive to think that organizational change itself will transform the INS," GAO's Hite said. Real transformation will require cultural change, business process change, systems change, performance management change and workforce change.

"These are things that give you the ability to make long-term change," he said. "And that will take many years."

***

Systems in Demand

Key information technology systems the Immigration and Naturalization Service needs include:

* Entry/exit system — Ordered by Congress and President Bush. Keeps track of foreign visitors who enter the United States legally and notifies INS if they do not depart when visas expire.

* Student-tracking system — Keeps track of the addresses and enrollment status of student visa holders to ensure that they are actually in the United States studying.

* Biometric identification systems — Required in legislation and endorsed by INS, biometric identifiers would make visas and passports harder to counterfeit and would reduce incorrect identification due to misspelled names.

* Remote sensors — With more than 6,000 miles of border to protect and a small force of agents, INS must depend heavily on remote cameras and other sensors to deter and detect illegal entries.

* Compatible databases — Systems operated by INS, the State Department, the FBI, the Customs Service and other agencies and systems within INS need to be able to share information so that criminals, terrorists and others can't slip past enforcement authorities.

* Basic case-tracking database for adjudication — "I can't tell you how often we call INS and reach someone who is good but [doesn't] know where the file is," immigration lawyer Elizabeth Stern said. INS needs a system that can retrieve documents when queried, she said.