For two dark years, the American auto industry has lived on the hope that a sales boom was just around the corner and the day was coming when consumers would flock back to the showrooms with open checkbooks.

A new, provocative analysis of the auto industry contends, however, that a boom simply isn't coming, because of an apparent fundamental change in consumers' demands for new cars.

A combination of high sticker prices, financing costs and gasoline prices has altered the rate at which consumers replace old cars with new ones or add to the numbers of cars they own, the study says.

Assuming sales do pick up next year and in 1983 from the deeply depressed levels of the past two years, the high-water mark still will be more than one million cars less than the industy is expecting, according to Data Resources Inc., a Cambridge, Mass., research group that performs complex computer analyses of the U.S. economy.

If true, the news could scarcely be worse for the American auto companies, which lost $4 billion last year overall and twice that on their North American operations.

Today, U.S. automakers seem unable to sell cars to a broad segment of the American population unless they offer big rebates. In February, March and August, car sales hit the 10.2 million-plus mark, on an annual, seasonally adjusted basis, thanks to rebates. Without rebates, the sales rate has hovered between 7 million and 9 million units, a reaction to the sharp increases in car prices and financing charges, according to consumer surveys. At rebate prices, most cars don't make a profit for the manufacturers, industry experts note.

In response, Chrysler Corp. has frozen prices for 1982 models at 1981-model levels on its most popular cars, and Ford Motor Co. is touting the economic prices of its older, mid-sized cars, and is offering rebates on other cars. Only General Motors Co. is taking a hard line on prices, with an average increase of 5.8 percent on 1982-model cars.

A surge in demand, permitting the companies to boost prices aggressively, would cure the industry's profit problems, but DRI doesn't see it. Instead, DRI projects total automobile sales, including imports, of 10.7 million units in 1983 compared with the depressed 9 million in each of the past two years. Even that recovery is far below what the industry says it needs, and what the more hopeful analysts have been predicting.

Value Line, an investor advisory service, for example, predicted last week that auto sales would rise to 11.5 million and 12 million in three to five years.

"Observers of this industry have been forecasting a turnaround for a year or more," Value Line's report said. "We believe that it may finally occur in 1982."

Donald E. Petersen, president of Ford, said recently that "there is so much pent-up demand in American and Europe that I foresee a couple of banner years between now and 1985 and 1986."

Robert Gough and Scott Mayfield of DRI contend, however, that the underlying demand for new cars has been fundamentally reduced. Between 1960 and 1974, this demand grew at an average annual rate of 2 percent. Since 1974, the growth rate has been only 1.2 percent a year, due primarily to the escalation in gasoline prices.

They calculated that the pent-up demand for cars among buyers who have delayed purchasing new models is now equal to about 2.3 million cars. This is equal to only half of the demand that would be expected based on trends before 1974.

However, other analysts believe that an easing in gasoline prices in the next few years could encourage a larger pick-up in car buying than the DRI estimates indicate.

"My numbers suggest a mild recovery in 1982, getting pretty strong in 1983 and 1984," said David Healy, auto analyst with the Drexel Burnham Lambert investment firm in New York. By the mid-1980s, an unusually large group of cars built between 1976 and 1979 will be candidates for replacement, and there is a limit to how long motorists will hold onto an old gas-guzzler, Healy said.