August 16, 1995
(Archived document, may contain errors)
August 16, 1995 WELFARE REFORM AND JOB TRAINING PROGRAMS: WHAT CONGRESS DOESN9T KNOW WILL COST TAXPAYERS ]BILLIONS
By Mark Wilson Rebecca Lukens Fellow in Labor Policy Consolidating federal job training programs and giving the funds to the states in the form of block grants is a major part of the Senate welfare reform debate. Because statistics indicate that more education leads to better-paying jobs, Congress and others assume that job training programs work. Unfortunately, they don't. Rather than focusing on who should have responsibility for job training programs, Congress must ask itself a fundamental question: Why should we continue to spend taxpayer money on programs that don't work? Job training programs, although well-intentioned, have not proven effective in raising the long- term hourly earnings of participants. They do not achieve this primary goal-better-paying jobs- because there is Ilittle, if anything, the government can do to alter the effect of neglecting the first 12 years of school. What is needed are more fundamental changes aimed at reducing illegitimacy and encouraging individuals to complete high school. Welfare reform that focuses on these two goals will reduce the demand for government job training programs that do not work. After 30 years, thousands of local job training programs, and hundreds of billions of dollars spent, only a few dozen job training programs have been subject to rigorous evaluation. The few rigorous studies that have been conducted generally fail to show that job training significantly increases long- term hourly earnings. Process-based performance measures traditionally have been the easier, but in- appropriate, method for determining whether these programs work. But performance measures can be manipulated easily to explain away poor results by those with an interest in holding on to federal job training money. If federal funding for job training is to continue, rigorous control group studies of these programs must be required! Without these studies, American taxpayers will never know whether they are funding effective programs or just throwing good money after bad. 2Y1#611.95
FEW STUDIES AND DISMAL RESULTS
Federal and state job training programs are intended to enable participants to build productive skills that will provide them with higher real wages. Only a few job training programs have ever been subject to rigorous evaluations, and the results show a record of failure.*5 As a report from the Secretary of Labor points out, "there are many areas where little thorough and reliable evaluation evidence is available." 4 The few solid studies that have been conducted generally fail to show any significant increase in the hourly earnings of participants directly related to job training.
A national controlled scientific study of the Title 11-A programs of the Job Training and Partnership Act (JTPA) reported these dismal findings: 5 X The program had no statistically significant effect on either the average earnings of young fe- males or their employment. X The program had a large (7.9 percent) negative effect on the earnings of young males and no effect on their employment. X The program had no statistically significant effect on the average earnings of adult men. X Although the overall program had a small (3.4 percent) effect on the average hourly wage of adult women, classroom training did not have any statistically significant effect. Further, training for those adult women that received welfare or were high school dropouts had no sta- tistically significant effect on earnings. X Immediate placement in a job did have a small effect on earnings, but not on higher hourly wages. The increase in earnings was due primarily to working more hours rather than to higher hourly wages. X Further, in 1991 the General Accounting Office found that the majority of service delivery areas underreported administrative expenditures, resulting in a misrepresentation of program costs and amounting to a circumvention of the statutory limitation placed on administrative costs. 6 A controlled scientific study of the Symmer Training and Employment Program (STEP) reported these disappointing findings: X After two to three years, STFP graduates did not have higher graduation rates or grades than the control group. X Graduates of STEP were no more likely to be working than the control group.
A scientific control group study of the Job Corps, one of the most expensive job training programs, hag never been conducted. However, a 1982 quasi-experimental study found these results: X Job Corps did not significantly help single mothers (the largest number of welfare recipients). X Job Corps did not affect the overall arrest rate. Further, a report by the Department of Labor's Inspector General found that only 17 percent of the Job Corps funding went to educational and vocational training, compared to 26 percent for ad- ministration.9 A study by the General Accounting Office also raised "serious concerns about the va- lidity of job placement information." 10 The GAO was unable to verify job placement for 15 percent of their Job Corps sample. After 30 years and billions of dollars spent, this program has never been subject to rigorous evaluation and only recently has begun to determine the employment status of Job Corps students. 11 A contr?lled scientific study of the Jobstart program also reported disappointing findings:1 X There was no significant increase in overall employment or earnings for the total four-year study period or for any individual year. X The program did not reduce welfare receipt. X Jobstart did not affect overall arrest rates. A controlled study of the Teen ie Parent Demonstration program reported only very It small improvements in earnings. X This demonstration project clearly indicates there are incentives associated with reducing wel- fare benefits to single mothers who do not stay in school or get a job. AFDC grants were reduced by $160 for those who chose not to participate. X However, even this significant penalty increased enrollment in school, job training, or em- ployment by only 13 percentage points and annual earnings by only $300. X There was no improvement in the poverty rate for participants. Controlled studies of the WIN and JOBS programs reported only very mall initial improvements in earnings, and these tend to "fade out" after five years: 14 X The main effect of these programs is to speed up the transition from welfare to work.
X There was only a small increase in average earnings per year, about $300 to $700, and after five years there was no difference in earnings between the control group and the group that participated in the program. X In most cases the increases in literac q skills and educational attainment did not translate into increased earnings or employment. A study by the General Accounting Office also found that the Department of Health and Human Services "does not know whether JOBS is reducing welfare dependency because it does not gather enough information on critical program outcomes." 16
After 30 years, thousands of local job training programs, and hundreds of billions of dollars spent, only a few dozen job training programs have been subject to rigorous evaluation. The few studies that have been conducted generally fail to show any significant long-term increase in hourly earn- ings directly related to job training. Performance measures and benchmarks are easy but unsatisfac- tory means of determining whether or not these programs work. They are easily manipulated and explained away. If federal funding for job training is to continue, rigorous control group studies of these programs must be required. One thing is clear from the few studies that have been done: Government can do little, if any- thing, to alter the effect of neglecting the first 12 years of school. The billions of dollars spent on government job training and public assistance programs are too much, too late. What is needed are fundamental changes aimed at reducing illegitimacy and encour- aging individuals to complete high school. In addition, parents need to make their children stay in school, do their homework, stay out of trouble, and graduate from high school. Welfare reform that focuses on these goals will reduce the demand for expensive government job training programs that do not work.
I Job placement assistance may speed up the transition from welfare to work and increase the number of hours worked per year, but that is entirely different from the primary purpose of training: better-paying jobs.
2 The job training block grant in the Work Opportunity Act of 1995 (S. 11-20) improves state accountability by requiring performance measures on the employment outcome of program participants, but it does not require control group evaluation studies.
3 The only method for evaluating job training programs effectively is to conduct an experimental design study that randomly assigns individuals either to a treatment group that can receive services from the program under study or to a control group that cannot, and then evaluate the outcomes. See Orley Ashenfelter, "Ibe Case for Evaluating Training Programs With Randomized Trials," Economics of Education Review, Vol. 6, No. 4 (1987).
4 What's Working (and What's Not) A Summary of Research on the Economic Impacts of Employment and Training Programs, U.S. Department of Labor, January 1995, p. 4.
5 The National JTPA Study: Title 11-A Impacts on Earnings and Employment at 18 Months, U.S. Department of Labor, Employment and Training Administration, 1993.
6 U.S. General Accounting Office, Job Training Partnership Act: Inadequate Oversight Leaves Program Vulnerable to Waste, Abuse, and Mismanagement, GAO/HRD-91-97, July 30,199 1.
7 Jean Grossman and Cynthia Sipe, "rbe Long-Term impacts of the Summer Training and Education Program," Public/Private Ventures, 1992.
8 Charles Mallar et al., Third Follow-up Report of the Evaluation of the Economic Impact of the Job Corps Program, Mathernatica Policy Research, 1982.
9 Semiannual Report, Office of the Inspector General, U.S. Department of Labor, April I - September 30, 199 1, P. 19.
10 U.S. General Accounting Office, JOB CORPS High Costs and Mixed Results Raise Questions About Program's Effectiveness, GAO/HEHS-95-180, June 30,1995.
11 As of July 1, 1995, the Labor Departments regional office verification system will be replaced by a follow-up survey of job placements to determine employment status 13 weeks after students leave the program.
12 George Cave et al., "JOBSTART: Final Report on a Program for High School Dropouts," Manpower Demonstration Research Corporation, October 1993.
13 Rebecca Maynard, "Building Self-Sufficiency Among Welfare-Dependent Teen Parents: Lessons from the Teenage Parent Demonstration," Mathernatica Policy Research, 1993.
14 Daniel Friedlander and Gayle Hamilton, forthcoming, Russell Sage Foundation.
15 Karin Martinson and Daniel Friedlander, "GAIN: Basic Education in aWelfare-to-Work Program," Manpower Demonstration Research Corporation, January 1994.
16 U.S. General Accounting Office, WeVare to Work, Measuring Outcomesfor JOBS Participants, GAO/HEHS-95-86, April 17,1995.