industrial area (23.3 percent), followed by internal support services (13.6 percent).

--Unit labor costs. The most important story in terms
of budgetary impact is measure in the cost per unit
or work produced. Productivity gains have sharply re-
duced the impact of wage increases and inflation in
the past five years, as the following data show, com-
paring fiscal year 1971 to 1967.

Our research continued by selecting seven functional subdivisions within these three broad groupings.1 Ex

hibit 2-9 shows the results within the three broad categories; there are sharp departures from the general pattern. The following highlights are noted:

See appendix G.

81-339 O 72-3

1. The public service "operating" activities, with a productivity gain of only 3.4 percent, are dominated by the Postal Service. However, another activity of major size is the health services--over 200,000 manyears--where a flat productivityl existed during fiscal years 1967-71 period. This underscores the need for a continued vigorous search for laborsaving break-throughs in these institutions.

2. The public service "processing type" activities, by contrast, show major gains in output--21.4 percent -with less than half this increase (5.3 percent) in man-years. This reflects the impact of automation and systems improvements on mass processing functions, such as Social Security and other activities with transaction-type workloads.

3. Support services of the "management type" have shown a small decline in workload, but a favorable productivity ratio has resulted from proportionately larger personnel reductions.

4. "Procurement and supply" activities have shown trends similar to the management services.

5. "Maintenance" of bases, buildings, and equipment shows a large increase in workload with a stable personnel manning. This produced a significant gain in productivity.

6. The "major overhaul" segment of the industrial category--highly defense related--reflects increases in workload, with a downward trend in man-years. This is an area which, if properly supported by competent industrial engineering and timely capital investments, should continue to show significant productivity gains.

7. Finally, the area of "manufacturing", with the smallest total man-years (19,000), has the largest gains in productivity (28.8 percent). The efficiency of such manufacturing enterprises as the Bureau of Engraving and Printing and the Mint accounts for this outstanding performance.

It must be stressed that the above observations are but starting points for further studies which should be carried

1The fiscal year 1971 index for the Health Services represented in our sample was computed at 101.2 percent of 1967.

on in individual agencies. These preliminary observations would be incomplete without some attempt to estimate future trends.

What Do the Data Suggest As To
Future Trends in Productivity?

The joint team believes the following trends may occur in fiscal year 1972 and 1973:

--Defense workloads will continue to drop. While substantial civilian personnel reductions were achieved in fiscal year 1970 and fiscal year 1971, this rate may be less rapid in the future. Also, many installations which have been fully utilized during the Vietnam period may be less heavily loaded. These factors may have a downward influence on Federal productivity indices in fiscal years 1972 and 1973.

--The Postal Service has shown a steady workload in-
crease in each of the past four years, with a level-
ing off of man-years between fiscal years 1970-71.
If one assumes a further workload growth with little
gain in man-years, this would have an upward influ-
ence on Federal productivity.

--With respect to the civilian agencies, it is assumed
that workloads in many public service programs
(health, education, environmental, etc.) will grow,
while employment levels will be held even or perhaps
show a small decrease under the President's personnel
reduction program.
Such trends would boost the pro-
ductivity indices.

The assumptions and judgments discussed above, suggest that the indices for fiscal year 1972 should show some growth, although perhaps at a slower rate than 1967--1971.

This ability to speculate on potential trends is one of the values of this measurement tool. This capability should stimulate managers to ask sharper questions and prepare more effective advance plans. This is particularly true, as discussed in the next chapter, of the need for timely planning of productivity enhancing capital investments.

Summary of Findings and

Proposed Next Steps

1. While continuing refinements will be required for at least one more year, we believe that a technique has

been developed which can measure trends in productivity of 55 to 60 percent of Federal civilian employees -- from year to year on a consistent basis.

2. Perpetuation of this technique will require the issuance of an OMB data call requesting all departments and agencies (of 200 or more employees) to submit the required data--shortly after the conclusion of each fiscal year--to a central agency for statistical aggregation and index construction. We recommend that the Bureau of Labor Statistics (BLS) evaluate the technique developed, and the quality of data available. If it believes that they meet acceptable standards then BLS should propose plans--including the resources required--for assuming this responsibility on a permanent basis. In the meantime, the data gathering and analysis should be conducted by the joint team with the aid of BLS.

3. Regardless of whether BLS becomes the central collection and publishing agency--it is recommended that OMB, CSC, and GAO continue this joint project for at least one more year with the following principal ob· jectives:

a. To work cooperatively with all agencies to increase the coverage thus far achieved to the maximum extent practical (an additional 100,000 manyears or more is estimated). In those areas where quantitative productivity measures are not feasible, studies should be conducted by the joint team of other ways of assessing productivity and performance trends. Such assessments might be used as supplementary data in the preparation of periodic reports to Congress and the public on total Federal productivity.

b. To improve the quality of the output measures, and to experiment with techniques of reflecting quality changes in computing the indices.

c. To develop additional or alternative functional indices for programs and activities where comparison among Federal agencies--and between the Federal, State, local, and private sectors--is possible and useful. The desires and needs of prospective users both in and outside of Government will be sought.