The jobless recovery enigma remains largely unsolved. As a special case of broader unemployment, the term “jobless recovery” describes an economic recovery where output recovers—and even expands—yet employment growth remains anemic. While the effects of these prolonged recoveries are significant—from increased crime to a lifetime reduction in wages—they are not well understood. Building on the insights of labor market matching models that incorporate heterogeneity among workers, this paper sheds light on jobless recoveries, developing a first-of-its-kind index of human capital heterogeneity for the unemployed, and testing that index using of a Structural Vector Autoregression. I demonstrate that the extent to which unemployed human capital is heterogeneous and specific, rather than homogeneous and general, plays a key and under-appreciated role in the labor market; increases in human capital heterogeneity can account for between one-quarter to three-quarters of the joblessness of the past three recoveries in the pre-COVID era.