What Works at Scale? A Framework to Scale Up Workforce Development Programs
Federal Reserve Bank of Atlanta
Community and Economic Development Department
Discussion Paper 2019-1
Workforce development policymakers have access to a growing catalog of training programs evaluated with rigorous randomized controlled trials. This evidence base identifies programs that work in specific geographic and temporal contexts but may not necessarily work in other contexts or at a scale sufficient to meet regional workforce needs. The author examines a sample of recent randomized controlled trials of workforce development programs and reports to what extent this body of evidence informs policymakers about what works at scale. The author finds that most programs are implemented at a small scale, use nonrandom samples from the population of interest, and are concentrated in the most populous urban areas and U.S. states. The author then discusses a method to help state and local policymakers, technical colleges, training providers, and other workforce development organizations adopt evidence-based policies in their local contexts and at scale. The two-step method includes a check on the assumptions in a program’s theory of change and an assessment of the sensitivity of projected results to violations in assumptions such as program completion rates. The author provides an example of the method applied to a hypothetical metropolitan area that seeks to adopt an evidence-based training program for youth with barriers to employment.
JEL classification: I38, J08, J24
Key words: workforce development, human capital, skills, provision and effects of welfare programs
The author would like to thank Ann Carpenter, Karen Leone de Nie, William Mabe, John Robertson, and Nisha Sutaria. The views expressed here are the author’s and not necessarily those of the Federal Reserve Bank of Atlanta or the Federal Reserve System. Any remaining errors are the author’s responsibility.
Comments and suggestions to the author are welcome at firstname.lastname@example.org.