Private shipyards consistently submit unrealistically low bids to win Navy ship repair jobs and then end up charging more than the Navy's original estimates of what the jobs should cost, a recent study has found.

The General Accounting Office examined 75 "fixed-price" contracts and found that all 75 ended up costing more than the fixed price agreed to, by an average of 63 percent. The study found that shipyards consistently underbid the Navy's internal estimate when competing for jobs with other shipyards, but consistently charged more than Navy estimates for unexpected "growth work" and "new work" once the ships were in drydock and competition was no longer a factor.

The GAO study appears to show that a chronic management problem, which Navy officials had said was almost solved, still plagues the Navy's private repair business. A related Navy study, which the service made available last week, argues that the problem is no longer as serious as the GAO makes it look. While not disputing the GAO figures, the Navy study said that during the period examined by both studies -- 1982 through 1985 -- the Navy record steadily improved.

The Navy study acknowledges that shipyards from Brooklyn to San Diego frequently "buy in" to contracts -- or, in the jargon, "low-ball" them -- and then attempt to "get well" on "growth work" and "new work." The Navy study attributes the prevalence of low-balling to the decline of the commercial ship repair and construction industry.

"Shipyards are looking to Navy repair work as the only alternative," the Navy explained. "There is a limited number of Navy overhauls . . . . The combined impacts of these forces have led to a very intense market."

But the Navy also concluded that low balling is not very harmful, because most overhaul jobs do not end up costing much more than the Navy estimated they should. In arriving at that conclusion, the Navy study excluded what it calls "new work" from calculations.

A Navy official explained the terminology by using the analogy of a garage. If a mechanic agrees to fix a transmission and then, once the car is in the shop, increases the price, that is "growth work." But if the mechanic then says that the brakes are shot, too, that would constitute "new work," the official said.

The GAO included both growth work and new work in its analysis and found that the 75 "fixed-price" contracts, negotiated for $594 million, eventually cost the Navy $967 million, an increase of 63 percent.

On the original contracts, the shipyards bid an average of 31 percent less than the Navy's estimates of what the jobs would cost. But on additional work discovered after the ship entered drydock, prices averaged 27 percent more than what the Navy thought was reasonable.

"A primary cause of the higher prices," the GAO said it was told, "was that modifications were negotiated on a sole-source basis with the original contractors and were not negotiated competitively."

The Navy study added, "Commonly, in order to 'get well,' the contractor attempts to overbid growth work or argue that work was not well defined in the original bid package."

The Navy said that some growth work should be expected, because not all problems can be identified until a contract is awarded and a shipyard can examine a vessel thoroughly. But growth work beyond 20 percent, the Navy said, is not considered reasonable.

Everett Pyatt, assistant Navy secretary for shipbuilding and logistics, said in an interview last summer that the Navy had taken large strides toward controlling the problem. But Pyatt, who was unavailable for interviews, also acknowledged that the problem would never be fully solved.

"In a deal like that, there's so much open-and-inspect work -- that's the ideal case to get ripped off," he said then. "I don't care whether it's your car, your house or your ship. Once you're in the shop, you're in the same situation."