persist. Since neither are known in this case, table 3 illustrates the situation with three alternative discount rates and durations of benefits. I leave it to the reader to choose the ratio between the costs and the benefits he prefers among these alternative illustrations. Annual Income by Enrollee Characteristics - Manpower program evaluators have always been plagued with the nagging probability that in less than full employment, any observed gains for enrollees were won at the expense of others who would have obtained those jobs and earnings in the absence of the program. Public expenditures on behalf of specific individuals or groups are generally accepted when one of three conditions prevail: 1) When the expenditure results in an increase in the gross national product sufficient that the entire society can have a higher disposable incore despite the expenditure; 2) When the expenditure results in a sufficient decrease in social costs that the same is true; and, 3) When the public has decided, through its representatives, to redistribute income in favor of that group. The GNP impact cannot be measured for MDTA, since it would be necessary to assume that the jobs obtained by the MDTA enrollees would have otherwise remained idle and that the production they were responsible for would not have occurred. There is also no data available to support the second justification. However, it is declared publi policy to use manpower and other programs to redistribute income in favor of those defined as disadvantaged; that is, those who are: 1. Both poor and without satisfactory employment; and Eligibility for MDTA requires only unemployment or underemployment. The disadvantaged are to receive priority to the extent that 'sixty-five percent of the overall enrollment should meet the designated criteria. However, the target need not be applied to any particular individual or project. The remaining thirty-five percent is expected to reduce labor shortages, an effect which could be expected to increase the GNP if it could be shown (a) that labor shortages existed and (b) that IDTA was training for them. The two objectives need not be mutually exclusive since disadvantaged persons can be trained to fill labor shortages. A vital test of MDTA's validity, therefore, is the proportion of disadvantaged among its enrollees, even if there is no assurance that others equally disadvantaged would not have obtained the jobs. Considering family The follow-up study produced only personal income data for the pre-training period, while the poverty criteria are derived for families. The personal income of non-family heads cannot be assumed to represent those of the family but the personal incomes of heads of households should be a reasonable approximation. size, assuming all household heads to be non-rural and under 65 (true of almost all), and assuming household heads to have been typical of the whole sample, sixty-nine percent of the Institutional enrollees and fifty-six percent of those in OJT net the definition of disadvantaged. ore than men By various characteristics, women gained slightly more from institutional enrollment with men gaining slightly more in OJT. Whites generally made greater gains than blacks with Spanish-Americans consistently making the greatest gains of any ethnic groups. The Impact of Program Mix Having looked at "who gains most", what can be learned of the relative contributions of alternative mixes of services? Originally, MDTA projects were organized in class size groups and were held in public schools during off-hours or in rental facilities. Subsequently, skills centers were established to provide a range of training offerings accompanied by on-site orientation, counseling, placement and supportive services. More recently, individual enrollees have been referred to on-going school courses with their tuition paid by MDTA. All three types now coexist and their relative effectiveness is of interest. There are significant differences among the income impacts of the three program types according to the measure used. The median income gain for individual referral enrollees was $2,380, compared to $1,980 for skils center enrollees and $1,860 for class size projects (table f Of those amounts, $961 for the skill center enrollees, $1,902 for the individual referrals and $1,054 for group-sized projects were accounted for by the employment of those who held no jobs in the pre-training period. Of the remaining portion of the income gains, $621 for the skill centers, $328 for individual referrals and $524 for class-sized projects were attributable to improved employment stability for those already in the labor force. The remainder, $398 for skill centers, $150 for individual referrals and 3242 for class-sized training were the consequences of wage increases. Apparently individual referral found heavy use and was effective for bringing into the labor force those who were previously outside it. Skill centers were most effective for adding to the employment stability of those already in the labor force and for increasing wage rates. This squares with the |