SHOULD PASSIVE restraints--air bags or auto matic seat belts, either with or without devices that make them detachable--be required in cars? A 1966 law gives the National Highway Traffic Safety Administration the power to decide, and NHTSA has decided--often, though not always the same way, concerning passive restraints. It required them in 1972, modified its requirement in 1975, suspended it in 1976, issued a new requirement in 1977 and delayed its deadline in 1981. Later in 1981 NHTSA rescinded the passive restraint requirement altogether, and the devices have never been required in cars because of the lead times given manufacturers.

Now the U.S. Court of Appeals here has ruled that NHTSA acted arbitrarily and without factual basis when it rescinded the 1977 passive restraint rule. The court has required NHTSA to produce in 30 days a schedule that will show how it will analyze alternative passive restraint systems and come up with another decision.

The court in effect has asked NHTSA to come up with some good reasons for not enforcing the passive restraint rule Congress has left standing after repeated and prolonged consideration of this issue. The court's criticisms of NHTSA's rationale seem well taken. The agency provided no factual basis for its conclusion that detachable automatic seat belts would not be used significantly more often than the seat belts the government already requires; and it gave only the flimsiest of reasons for its refusal to require other forms of passive restraint. In other words, the executive branch did what Congress could easily have done but declined to do: change the basic thrust of an important law.

The Reagan administration advertised its initial delay and subsequent rescission of the passive restraint rule as one dose of medicine for the ailing auto industry. It is worth noting, however, that the industry had long notice of the requirement and had already spent much of the money required to meet it, and that at least one form of passive restraint--the automatic seat belt--would have added only minimally to auto prices.

We are by no means persuaded that the passive restraint rule is needed at all. Government has already provided most auto passengers with the means of protecting themselves, by requiring seat belts in cars, and we do not think government needs to go further by using the regulatory process to protect those who do not choose to protect themselves. We are also sympathetic to the arguments that this is not the time to pile more burdens on the auto makers or to do things that will raise even higher the price of new cars.

But this does seem to be a case where the Reagan administration has gone about the task of what it calls "regulatory relief" the wrong way. The principle that regulatory action should be predictable and not subject to sudden changes works against, not for, what the administration has done. NHTSA can now try to scuttle the passive restraint rule in a way that will pass muster with the courts, but the administration would be better off seeking straightforwardly from Congress the change in the law it wants.