AGRICULTURE SECRETARY Richard Lyng says the bottom-dwelling farm economy is looking up again. That's what agriculture secretaries are supposed to say as election years begin, and Mr. Lyng skated as quickly as he decently could over one of the main causes of the turnaround, which is that the government continues to spend more than $20 billion a year on farm income and price supports. Though down from the record levels of two years ago, these transfusions from the rest of us are still about seven times the annual average for the 1970s and more than a third of net farm income.

So the good news is to some extent a political and governmental artifact, but Mr. Lyng is also right. Markets are finally tightening. Prices, exports, net farm income, farmland values -- all are up or steadying, and debt and surpluses are down. That doesn't mean there are no more problems, just that the context has shifted. The 1985 farm bill was written in an agricultural depression; the issue was how to apportion failure. Price supports were too high; they and the production they were eliciting had to drop, but not so rapidly as to strand more people than necessary. The bill was a slow parachute.

Mr. Lyng says the policy has worked, which to some extent it has, though as ever in agriculture, the weather and exchange rates have mattered more. Now the issue has become the pleasanter one of how to apportion recovery. But even in this there will be losers. Two examples:

1) Marketing loans. These are an aggressive variation on the support pattern of the past. The way the older system works, the government guarantees the farmer a price. To help achieve that when markets are low, it buys his crop and stores it. The government thus supports the markets by holding crops in reserve. With marketing loans, the government also guarantees the farmer a price, but doesn't buy his crop. He sells that on the market for whatever it will bring, and the government pays him the difference. Instead of supporting the markets, this new arrangement floods them. It is a policy of subsidize-and-dump which seeks not very subtly to crush weaker foreign producers.

The marketing loans were tried for the first time in 1985 in rice and cotton. Both crops have now gone from surplus to strong demand. Proponents attribute this to the marketing loans, which -- though costly in the early years -- some enthusiasts now want to extend to other crops. In fact, though the marketing loans may have helped a little, the rice and cotton markets tightened mainly because of poor weather in Asia and other growing areas. Marketing loans are a form of agricultural imperialism; they would be the wrong way for this country to go.

2) ARPs. These are acreage reduction programs: a farmer who wants government supports has to promise to idle a certain percentage of his acreage in return. The percentages are now quite high. Between the ARPs and a conservation program begun in 1985, almost as much land is now idled as was set aside in 1983, when the government went to a payment-in-kind program, giving farmers grain from its bins if they would suspend a year's harvest. As markets firm, there will be pressure to reduce the set-asides. The pressure will come from consumers who want lower prices and from farmers who are afraid of ceding foreign markets; if we don't expand, others will.

The secretary is already having to pick his way among such pressures in rice and several other recovering crops. As with the marketing loans, so here: the need is for restraint. When the weather is right, world agricultural productive capacity continues to be greater than demand. There are too many farmers. It remains up to governments to civilize the competition among them