This longitudinal study examined the dynamic relationship between work engagement (vigour and dedication) and symptoms of anxiety and depression. A sample of 3475 respondents from eight different occupational groups (lawyers, physicians, nurses, teachers, church ministers, bus drivers, people working in advertising and people working in information technology) in Norway supplied data at two points in time with a 2-year time interval. The advantages of longitudinal design were utilized, including testing of reversed causation and controlling for unmeasured third variables. In general, the results showed that the hypothesized normal causal relationship was superior to a reversed causation model. In other words, this study supported the assumption that work engagement is more likely to be the antecedent for symptoms of depression and anxiety than the outcome. In particular, the vigour facet of work engagement provides lower levels of depression and anxiety 2 years later. However, additional analyses modelling unmeasured third variables indicate that unknown third variables may have created some spurious effects on the pattern of the observed relationship. Implications of the findings are discussed in the paper. Copyright © 2011 John Wiley & Sons, Ltd.