The American cow is causing problems in Washington.

The basic problem is simple enough. In 1977, Congress told dairy farmers to raise more cows. They did. But nobody wanted to buy the cows' milk. So the government did.

We are now heading toward the largest dairy surplus in history, and it's piling up in front of the Treasury. In 1981, the government will pay an estimated $1.6 billion so dairy farmers can produce surplus milk. If nothing is done, next year's costs will be even higher.

All this may seem too simple and silly to be true, but it's precisely what happened. The whole system of dairy price supports is so expensive and contorted that you might think Congress and the Reagan administration would rush to undo it.

Wrong. Although Congress is likely to modify the system, the prospective changes aren't nearly as radical as much of today's shrill rhetoric suggests.

Dairy price supports are not only a vivid illustration of how old programs have been overwhelmed by changing economics and politics, but also a reminder of the staying power of the status quo. Most constituencies aren't evil, and most members of Congress aren't eager to upset the old order for pure fun.

Much of the milk story, of course, has nothing to do with politics. There would be no problem at all if Americans were simply more fond of milk. In 1960, percapita milk consumption totaled about 33 gallons, more than double soft-drink consumption of 14 gallons. By 1979, Americans were selling 38 gallons of soft drinks compared with 28 gallons of milk. Beer, up about 33 percent since 1960, also had edged ahead of milk.

Nor has politics had much to do with the phenomenal transformation of American dairy farming. The average dairy cow produces almost three times as much milk as in 1925 and twice as much as in 1950. Better breeding, more feeding and better management account for most of the change. More improvements loom; milking three times daily instead of two may raise output again.

Dairy price supports haven't affected these trends much. The programs originated in the Depression, when government sought to protect farmers -- and many others -- from the unfettered free market. Preserving the samll family farm became another rationale. By that measure, the supports have not worked. The number of commercial dairy farms has dropped from 600,000 in 1950 to about 200,000 today.

But the program survives anyway. The government now pledges to buy certain milk products (dry milk, American cheese and butter) at official support prices to provide a floor under all private milk sales. And the support prices are tied to the so-called parity index of farmers' costs in the 1910-14 period. As the parity index rises, so does the support level.

If you ask what parity has to do with dairy farming, the answer is next to nothing. The index represents the costs of all farmers, not just dairy farmers. More important, the whole parity concept is flawed. Times have changed since 1910-14, and tying all farm prices to parity would create a self-perpetuating inflation machine. Any inflation would raise food prices immediately.

That's why Congress doesn't like grain-support prices to parity. But dairy is different, and Congress, listening to the dairy lobbies, made the difference worse in 1977. Instead of a yearly adjustment, Congress decided to raise the support price twice annually.

It couldn't have sent a clearer signal. Twice-annual adjustments made dairy farming highly profitable and, in addition, the parity index had the extra advantage of increasing more rapidly than the consumer price index and other inflation indicators. Farmers acted rationally.

For the first time since the mid 1950s, the nation's dairy herd increased. Given the rising productivity of cows and stagnant demand -- population growth and higher consumption of processed dairy products have roughly offset declining milk use -- there was no need for it. Most of the extra production predictably went into government stocks.

Just why there should be any official support program for milk isn't clear. There are no support programs for beef, poultry, pork or eggs. Although the prices of these foods fluctuate, no one argues that an official support program is essential to assure supplies.

The best justification for price supports is that they smooth excessive fluctuations in supply and prices. The government buys during periods of peak production to prevent a price collapse that would destroy so many producers that a price explosion would result later. But stabilizing prices doesn't requre the absurd automatic parity index and ought to involve much smaller government purchases than today: maybe 2 percent to 3 percent compared with the current 7 percent.

But neither the Reagan administration nor the Agriculture committees are considering such sweeping revisions. The administration did persuade Congress to forgo the 1981 semiannual adjustment on April 1. And it's pushing legislation that would allow the Agriculture secretary to set once-a-year support prices at 70 percent of parity rather than the current 80 percent.

That's a jolt in contrast with the past. As a practical matter, it may mean that the support price won't increase next October. But it still may require the Agriculture Department to make considerable surplus ll purchases. The Congressional Budget Office estimates $900 million worth.

What the current system has going for it is that it has been going for a very long time (since 1949). Plus, it's tough to get too mad at dairy farmers. As a dairy-state congressman put it: "They don't feel guilty about being farmers. They make a good, healthy product, and they work 365 days a year."