BACKGROUND: In a large body of previous research, cross-lagged panel models (CLPMs) have been used to provide empirical support for developmental models that posit a cascade from externalising-to-internalising problems. These developmental models, however, arguably refer to within-person processes whereas CLPMs provide a difficult-to-interpret blend of within- and between- person effects. METHODS: We used autoregressive latent trajectory models with structured residuals (ALT-SR) to evaluate whether there is evidence for externalising-to-internalising cascades at the within-person level when disaggregating between- and within- person effects. We used eight waves of data (age 7-15) from the Zurich Project on Social Development from Childhood to Adulthood (z-proso). RESULTS: ALT-SR fit better than the corresponding CLPMs. Using an ALT-SR, we found evidence for externalising-to-internalising cascades, consistent with previous CLPM studies. However, we also found some evidence for effects in the ALT-SR that were not apparent in the CLPM, including a negative effect of externalising on internalising problems in adolescence. In addition, a negative effect of internalising on externalising problems in adolescence was found in both the CLPM and ALT-SR. CONCLUSIONS: Within-person results were largely consistent with previous evidence from CLPMs; however, at the within-person level, externalising and internalising may negatively influence one another in adolescence.