INTRODUCTION

In July 1968, the Olympus Research Corporation contracted with the Manpower Administration of the Department of Labor to attempt to measure the total impact of all manpower programs in three contrasting metropolitan areas. In effect, what was pursued was the answer to the question, "In what way does each of the cities differ today from the situation in which it would have been had there never been any manpower programs?" The cities chosen for the study were the San Francisco Bay Area, with its complex and overlapping labor markets and political jurisdictions; Boston, large but unitary; and Denver, with its modest size and relatively simple

structure.

Every conceivable impact could not be measured, but the critical impacts seemed to be those on (1) the employment and earnings of the enrollees; (2) the local economy in terms of total expenditure and staff jobs; (3) the institutions of the community; and (4) the labor market. Then to give operational meaning to the impact measures, those manpower services accounting for the greatest positive impact were identified, and the implications of the findings for current policy issues were explored.

The study consisted of: (1) a one-year follow-up of a sample of 1709 individuals enrolled in that set of programs designed to improve basic employability; that is, those providing orientation, basic education, language facility, or job skills; (2) nearly three years of observation of the administrative capabilities of the various agencies operating within their community environments; (3) an inventory of the total expenditure and enrollments in manpower programs over four or more fiscal years; (4) an assessment of the extent to which staff jobs within the programs served as ladders for upward mobility out of the poverty community; and (5) a comparison of the magnitudes and occupational structure of enrollments with the size of labor forces and the structures of demand in each community.

The full, detailed report on the study comprises 24 chapters and approximately 1000 pages. A summary report of approximately 100 pages contains the essential information. This precis is just that--an exceedingly brief look at the high points of a lengthy study, hopefully serving to whet, rather than satiate, the appetite for further information.

THE EMPLOYMENT AND EARNINGS IMPACT

Table 1 summarizes the nature of the follow-up sample, and Table 2 does the same for its results. For purposes of the follow-up, the three metropolitan areas were treated as four cities--Boston, Denver, San Francisco, and Oakland. The essential finding was that across cities and across programs, the results were on the average positive and substantial.

It was unforeseen in planning the follow-up study that its initial interviews

would occur just as national unemployment rates began to rise and would continue in a steadily slackening labor market. Despite the unfavorable economic climate, the impact of program participation across programs, service functions, and cities was surprisingly positive.

Comparisons were made between a period of 36 months prior to enrollment (Period I), 12 months prior to enrollment (Period II), and the entire period between each enrollee's leaving the program and final interview (Period III)--a maximum of 12 months and an average of 9 months. Table 2 summarizes the results.

A more detailed analysis and further data display are found in the full report. This summary offers the main conclusions and sufficient data to support and illustrate them.

On the average across the four cities, the average hourly wage rate rose by $0.50 an hour between Periods I and III and $0.42 an hour between Periods II and III. The average hourly wage differences ranged from $0.30 an hour in Denver to $0.76 in San Francisco. Any citywide wage difference exceeding $0.12 an hour was statistically significant at the 1-percent significance level. Of course, wages were rising during the period between the first- and fourth-wave interviews. However, there were no minimum wage increases between the pre- and post-training periods and little unionization to push up the bottom of the wage structure within slackening labor markets. Since the midpoints of 36 months and 12 months are a year apart, the 12-cents-an-hour difference between the average hourly wage rates of Periods I and II may be an indication of the normal wage change in the tighter labor markets which prevailed between 1966 and 1969 and which included a substantial rise in the minimum wage. The normal increase during 1970 should have been no greater. Therefore, though the study included no control group to separate enrollment gains from those which would have occurred in the absence of program enrollment, there can be no question that the impact of the programs on the earnings of the enrollees was substantial and, in some cases, spectacular.

Improvement in average hourly wage rates measured only one dimension of the employment and earnings impact. The average enrollee who had work experience during Period I was employed 54.4 percent of the available time during that period. The average enrollee employed at any time during Period II had a job for 56.6 percent of the available time during that period. The average worker in the sample who sought and found work during Period III was employed 63.4 percent of the available time. A 4.5-percent increase is necessary for statistical significance at a 1-percent level. Thus, the improvements in employment stability were highly significant for San Francisco and Oakland. In Boston, the change in employment stability over Period I was significant; but that over Period II was at the margin of significance. The employment stability changes for Denver were not statistically significant, though the wage changes were.

The employment stability measure compared only those with some employment during the various periods. If those who had no work experience are included in each of the periods, a new measure which might be called employment intensity

69-110 O 71 33

emerges. The total groups were employed on the average 44 percent of the available weeks in Period I, 36 percent of the available weeks in Period II, and 48.1 percent of the available weeks in the post-training period. Improvement in employment intensity was significant for all the cities but Denver, but it is depressing to note that only in Boston was the full post-enrollment sample employed on the average more than one-half the time. As will be illustrated later, this is probably the contribution of WIN with its preponderance of female family heads on welfare who did not enter the labor force after training.

Income could improve from either an increase in wage rates or an increase in employment stability. Multiplying the two measures produces an income coefficient easily computed into annual income equivalents. For instance, if an individual worked 75 percent of the time at $2.00 an hour, his income would be the same as if he worked fulltime at $1.50 an hour. Multiplying the latter rate by 2000 hours would give the same annual income equivalent for either case.

The income difference between the income coefficients of Period I and Period III was $0.69 an hour, while that between Period II and Period III was $0.61. The average differences for Boston, San Francisco, and Oakland were all highly significant. Denver's income gains between Periods I and III were significant, but a slight drop in the income difference occurred between Periods II and III because pre-enrollment employment stability was enough higher to offset the hourly wage

gain.

Translated into annual income equivalents, the average enrollee across the four cities would have gained $1380 a year over Period I and $1220 a year over Period II. Transformed into percentages, this would have meant a 40-percent increase over Period I and a 30-percent increase in income over Period II. However, the average enrollee who worked following training was still earning only at a rate of $3000 per year. Poverty had been made substantially more comfortable, but not eliminated.

The enrollees, the program mix, and the results differed substantially for the four cities. Table 3 summarizes trainee characteristics by city.

Of those with pre- and post-training employment, the average hourly wage increase over Periods I and II, respectively, were $0.50 and $0.46 in Boston, $0.35 and $0.30 in Denver, $0.76 and $0.58 in San Francisco, and $0.40 and $0.33 in Oakland. Employment stability also improved substantially from 62.9 percent and 68.7 percent for the first two periods to 72.6 percent for the post-training period in Boston. Similar calculations for the other cities show 54.4 percent, 61.8 percent, and 57.2 percent, respectively, for Denver; 55 percent, 47 percent, and 63.3 percent for San Francisco; and 45.3 percent, 49.4 percent, and 60.3 percent for Oakland.

Comparisons of the wage and employment stability measures can again be combined into a measure of income increases and equivalent annual incomes. For