to $1.8 billion over a four-year period), the food stamp program (still expanding and estimated to cost $193 million in fiscal 1968), the manpower training program of the Department of Labor (for which spending is estimated at $295 million in the current year), and the air pollution control program.

The recent increases in spending for a series of new health research and services activities largely reflect the establishment of several new national institutes of health, and the initiation or expansion of health facilities and training

programs.

In addition to these 21 new programs, three more were established in fiscal 1967, and the 1968 budget proposes the enactment of an equal number. Of these, by far the most significant is the supplementary health insurance program under which the Federal government matches the contributions of individuals to pay for additional medical care. Federal expenditures for this supplementary medical care plan will approach $1 billion in each of the fiscal years 1967 and 1968, and can be expected to rise in future years.

A major increase in expenditures also results from the addition of the so-called "medicaid" program, providing grants to the states for medical assistance to the needy. Expenditures for this and other public assistance grants will increase by about $1.1 billion between fiscal 1965

and 1968.

Public assistance grant programs have an interesting history. Expenditures for such grants have increased from $1.5 billion in 1956 to an estimated $4.2 billion the current year. When the Social Security Act was originally enacted in the mid-1930's, it was predicted that as more of the aged became eligible for the old

age and survivors' benefits, public assistance costs would decrease. Instead, repeated actions by the Congress extending coverage of the program, increasing the benefits, and establishing new categories of public assistance grants have resulted in substantial increases in Federal expenditures for such grants.

While new programs in health, labor, and welfare created over the period of this study add up to significant totals in 1968, legislative extensions of older established programs have had an almost equal expenditure effect. In total, administrative budget outlays for health, labor, and welfare grew from $2.2 billion in fiscal 1955 to an estimated $11.3 billion in 1968.

Beginning in fiscal 1956 and through 1966, there were established 15 new Federal programs in the field of education. By fiscal 1962, under the impetus of new programs established under the National Defense Education and other acts, and broadened aid to “Federally affected" school districts, Federal spending for education exceeded $1 billion for the first time.

The major new programs were the three added in fiscal 1959 under the National Defense Education Act, and the elementary and secondary and higher education assistance programs, among 5 new enactments in 1966.

With the advent of the new elementary-secondary and higher education aid programs, Federal education expenditures under the new programs covered in this analysis more than tripled in fiscal 1966, totaling $1.7 billion in that year.

The cost of the new programs was $2.1 billion for fiscal 1967 and is projected at $1.5 billion for fiscal 1968. Net expenditures for all Federal activities included under the standard classification of education rose from $1.5 billion in fiscal 1965 to $2.8 billion in 1966.

The totals budgeted for educational programs in fiscal 1967 and 1968 ($3.3 and $2.8 billion, respectively) are distorted by the effects of scheduled sales of participations in pools of Federally held college housing loans. Such sales are expected to exceed new expenditures under that program in both years, with the excess being counted in the budget as a reduction in expenditures. If the effects of these participation sales are eliminated, education expenditures would be $0.3 billion higher than shown in fiscal 1967, and $1.3 billion higher in the current fiscal year.

Is there a characteristic pattern of growth of a new Federal program after it is initially put into operation? The actual dollar amounts expended for each program, as shown in the Appendix tables, reveal by simple inspection that virtually all of the individual programs experienced some growth after their beginning years. Further, the estimated 1968 cost of the programs adopted through fiscal 1966 amounted to $14.8 billion, as noted above, almost seven times the collective first-year costs for this group of programs (Table 1).

These comparisions, however, do not

provide a basis for generalization regarding the typical growth pattern, if any, for the individual programs. There are at least two reasons why this is so: (1) the estimated 1968 costs apply to programs which, by the end of 1966, had been in

more successive years. These measures, based on first-year costs as 100 percent, were grouped in a frequency distribution; and median index numbers were computed for the costs in each successive year. The results appear in Table 4, and in Chart 2.

The average (median) index of change among the programs so analyzed increased four-fifths in the second year from the initial year's base. In part this reflects the fact that programs do not usually operate for a full 12 months, or on the intended scope, in the first fiscal time is required for a new program to beyear in which they begin, as some lead

come operational. The increase in thirdyear costs from the second year was about three-fifths. In the fourth and fifth years the average cost index rose more

1 to 11 years, and thus all had not had equal exposure to potential sources of growth through time; and (2) the totals, though important in themselves, may reflect unduly the weight of a relatively small number of individual programs of large initial size or with large dollar increments over their lifetime.

To arrive at a more representative picture of the typical growth of individual programs, a series of annual index numbers was computed for each program with observations available over two or

and fourth years, and another one-fourth in the fifth year. In the sixth year, however, it jumped by more than one-half. (Increases continued in the successive

years; however, the number of observations thins out sharply by the seventh year, to only 15, so that the results are not very meaningful.) By the sixth year, the median index of program costs was over 6 times the first-year index and 31⁄2 times the second-year index.

The general pattern characterizing the growth of these new programs is this:

sharp increases in the first two years as the programs get into fuller operation, relatively modest increases in the third and fourth years, followed by a steep jump of the sort depicting major expansion or legislative extension of the program.

The distributions presented in Table 4 give further insight into the spread of program changes over the life span of individual programs. A small number declined in cost from their initial expenditure levels, and only one remained unchanged. In the early years, the program growth patterns tended to be grouped in brackets registering gains of less than 100 percent (from the initial year), but even in these years many programs rose to multiples of 2 to 10 or more times their first-year cost levels. With successive years, an increasing portion of the programs shifted into the growth categories

depicting ever larger gains. By the sixth year, one-fourth of the programs were ten or more times their first-year size.

Since final cost data were available

through the year 1966 only, the analysis of growth patterns includes the new programs which were established through fiscal year 1965 only, and only the secondyear patterns of growth for the new 1965 proprograms. As noted elsewhere, these grams as well as those enacted in 1967 and proposed in 1968-differ significantly in scope from most earlier programs. Some represent excursions into new fields of activity traditionally financed from non-Federal sources.

Whether the growth patterns of these programs can be applied to a broader sample of programs over time is moot. Further, additional analysis is needed of the patterns of change in relation to the