Objectives Workplace bullying has been suggested to increase symptoms of anxiety. A reverse relationship has also been proposed. However, so far only few earlier studies have investigated this topic and the reported associa-tions might partly be explained by unmeasured individual characteristics. In this study, we aim to examine the temporality and directionality between workplace bullying and anxiety symptoms, taking time-invariant charac-teristics into account. Furthermore, we aim to examine whether leadership quality modifies these associations. Methods We included 13 491 individuals from two nationwide cohort studies in Sweden and Denmark. Using cross-lagged structural equation models (SEM) and dynamic panel models with fixed effects, we examined con-temporaneous and lagged associations between self-reported workplace bullying and anxiety. Cohort-specific results were estimated and combined using fixed-effect meta-analysis. Results The cross-lagged SEM models supported contemporaneous and lagged relationships in both directions (from workplace bullying to symptoms of anxiety and vice versa). In contrast, only contemporaneous relation-ships remained statistically significant and of considerable magnitude in the dynamic panel models with fixed effects. Specifically, exposure to workplace bullying was related to a concurrent increase in anxiety symptoms (b=0.61, 95% confidence interval 0.32-0.90). No support of interaction with leadership quality was found. Conclusions The results indicate that onset of workplace bullying is associated with an immediate or short-term increase in anxiety symptoms. This study provides novel insights regarding temporal aspects and causal inference of the bullying-anxiety relationship useful for managing psychological hazards and preventing mental illness at work.