THE REAGAN administration's commitment to run federal programs more efficiently can easily run counter to its commitment to reduce federal paper work. In the case of the new job-training program being planned by the Department of Labor, the paper-work controllers are currently threatening to derail a carefully planned system for finding out whether the program really works.

The administration hopes to compensate for some of the cutbacks in job programs by improving the training programs that, under CETA, were run primarily by local governments. Under the Job Training Partnership Act passed by Congress last year, the CETA program is being transformed into a new system in which state governments and the private sector will have a stronger role. The Labor Department wants to continue giving states and localities a good deal of flexibility in deciding what programs suit their areas. But the department also recognizes that if the new system is to avoid the charges of waste and misuse that plagued CETA, it will need to know how well each local area is running its programs.

States and localities--even those that run excellent programs--have traditionally resisted federal efforts to monitor their performance and enforce standards. Their best argument is that simple success measures, such as the number of trainees who subsequently find jobs, don't take account of important differences. A program in an area with a low unemployment rate, for example, will find it much easier to place trainees. So will a program that steers away from school dropouts and other hard- to-place people, even though those are the kind of people most in need of help.

To answer these objections the department consulted with the various groups involved in running the program--states, localities, business and labor --and came up with a reporting system that, amazingly enough, has won strong support from every one of them. No data are required that would not be routinely collected by any minimally well-run training operation--if anything, the information collected on types of training and placement activity should be expanded. Yet the information is sufficient to let the department determine whether, taking into account differences among trainees and areas, a program is working well--something that a more limited sampling of programs couldn't do.

The red-tape fighters in the Office of Management and Budget, however, are resisting the system on grounds that it is an undue burden on states and localities and inconsistent with the notion of "block grants." That's a misreading on two counts. States and localities recognize that they need to monitor their programs anyway and that it's just a question of whether that information will be useful in comparisons with other areas. Nor should federal taxpayers be asked to support local programs without the assurance that the people who run them will be held accountable for their performance.