Judging by recent comments, Defense Secretary William Cohen appears to be leaning toward a recommendation that President Clinton take the first step toward deploying a national missile defense system.

There is little doubt Cohen believes a shield can be built; the issue may come down to cost and timing.

Cohen is likely to take another week or two to consider many complex aspects of this decision, from the urgency of a missile threat to whether an anti-missile shield is affordable.

"I frankly think that if you had an ICBM (Intercontinental Ballistic Missile) that landed in New York or Washington or San Francisco, Los Angeles or Tampa, the damage would be catastrophic, and the amount of money ... would be rather insignificant," Cohen said July 6.

The proposed missile defense system is projected to cost $60 billion.

As recently as Wednesday, in remarks to reporters in his Pentagon office, Cohen said he had not made up his mind what to recommend.

His advice will be an important factor in Clinton's decision, but the president also will take into account views of the rest of his national security team, including Secretary of State Madeleine Albright.

Cohen has noted that even though Clinton must decide whether to start the deployment process, his successor will be able to stop it, modify the plan or take an altogether different approach.

The initial construction work on an X-band radar in Alaska cannot begin before summer 2001.

The question for Clinton is whether to give the Pentagon the go-ahead to award construction contracts to let site preparation begin next year. If Clinton decides not to go forward, he will have eliminated the possibility of having a national missile defense ready by 2005, as now planned.

Whether that target remains realistic would seem the first question Cohen must consider.

His recent comments give indications of doubt. In testimony Tuesday before the Senate Armed Services Committee, Cohen noted the failure of the past two flight tests of the "smart rock," or missile interceptor, that is being developed to shoot down incoming warheads.

"Perhaps it has called into question the realism of the date," Cohen said. But he added he considers 2005 the target "we ought to continue to focus on" in order to keep the work — both the testing and construction of the anti-missile system — moving as fast as possible.

Even if Cohen were to determine that 2006 or 2007 is a more reasonable target date, he might want to proceed immediately with the first phase of construction in Alaska to lengthen the odds of meeting that timetable.

The weather on Shemya Island, in Alaska's Aleutian chain, where the Pentagon plans to build the powerful new radar, is so foul that construction is possible only a few weeks each summer. The radar is needed to help the missile interceptors find their targets and to distinguish decoys.

Cohen said Wednesday that Gen. Larry Welch, head of a panel of experts advising the Pentagon on the missile defense project, has said Shemya's weather might unavoidably delay construction past 2005.

During the 18 months since he altered the 2003 target deployment year to 2005, Cohen often has said he believes a national missile defense is justified by an emerging threat of attack by North Korea or possibly Iran or Iraq. The urgency of the threat is among Clinton's principal considerations.

"There is a threat, and the threat is growing," Cohen said on Jan. 20, 1999.

Clinton seems to share that view.

"Is there a threat which is new and different?" Clinton asked at a May 31 news conference in Lisbon, Portugal. "The answer to that, it seems to me, is plainly, yes, there is, and there will be one."

Another major factor in Cohen's decision is the technical feasibility of the missile defense system under development. Many critics in Congress and elsewhere contend it cannot be made to work reliably.

Cohen believes it can.

"We believe that the testing to date demonstrates the validity that we are close to having a technology that can, in fact, defeat a few dozen missiles fired by a rogue state, and that's the criterion that I will look at," he said June 12. That was before the latest test failure, July 7, the second consecutive setback.

In his testimony Tuesday, Cohen acknowledged that the two failures leave in doubt the feasibility question. He quickly added: "I believe that the trend is such that these are problems that are correctable."