The effects of mediums of instruction on educational- and labor market outcomes: Evidence from Malaysia

We examine the effects of having English as a medium of instruction on labor market outcomes later in life. We exploit an exogenous variation in mediums of instruction induced by the government of Malaysia’s decision to discontinue English-medium public schools and phase them out with Malay-medium public schools in 1970, which fits a fuzzy regression discontinuity design. We find some evidence that having English as a medium of instruction improves labor market outcomes. We explore some mechanisms through which mediums of instruction matter: We find having English as a medium of instruction improves English proficiency, especially reading and writing skills, and increases educational attainment, which in turn increase earnings and improve employability. The evidence is, however, rather weak, if we use robust data-driven inferences in the regression discontinuity design.