LAST YEAR, the administration focused its efforts on eliminating those features of welfare law that provided encouragement for work by gradually reducing welfare benefits as earnings increase. This year, the administration not only proposes more of the same treatment for the working poor, but has abandoned its commitment to protecting the "totally dependent" as well.

A new study by the Center for the Study of Social Policy reveals that under the latest plan, no recipient, however poor, would be immune from loss. Benefits would be cut even in states such as Alabama, Mississippi and Texas, where families unable to work must subsist on a total income--including welfare, food stamps and fuel aid--of only half the poverty level. On average, nationwide total welfare income would be cut to 72 percent of the poverty level.

What about the families that try to help themselves by working? Back in the "old days" of welfare debate, it was thought that people who took the sort of ill-paid, periodic jobs that are the welfare recipient's normal lot should keep at least a third--and preferably more--of their earnings for their efforts. In the Reagan scheme, however, the question is not how much better off working recipients should be, but how much worse off. In California, for example, a working recipient could lose over 30 cents in net income for each additional dollar earned.

The fact that welfare families--even before the Reagan cutbacks--received little net return from working has been shown to encourage not only dependency but family breakup and welfare cheating as well. And because, over the course of a year, most welfare families have at least one member working, reduced work incentives can have a pervasive effect on perpetuating dependency. These facts are apparently not unknown within the administration.

The recently issued Economic Report of the president observes that welfare recipients "may well face the highest marginal tax rates of any members of our society" and "not surprisingly, therefore, relatively few beneficiaries . . . work." This being the case, the report concludes, improved work incentives for these people "might generate a sufficiently large addition to their labor supply to pay for themselves." How true.