What is the value added by caseworkers?

This paper examines the efficacy of caseworkers in allocating individuals to government programs and to services within those programs. We investigate caseworker allocation of unemployed individuals to subprograms within Swiss active labour market policy in 1998. Our analysis compares the caseworker allocation to alternatives including random assignment to services and allocation via statistical treatment rules based on observable participant characteristics. Using unusually informative administrative data, we find that Swiss caseworkers obtain roughly the same post-program employment rate as random allocation to services, while statistical treatment rules, even when subject to capacity constraints, do substantially better.