Taken together, prior publications on the association between symptoms of depression and anxiety and contact with the criminal justice system (CJS) suggest a bi-directional relationship, but all the studies only focus on one direction in this relationship. To examine, in longitudinally collected data, period-specific within-individual change in anxiety and depression measures preceding arrest measurement and, separately, following arrest measurement. Data were obtained from the National Longitudinal Study of Youth 1997, a nationally representative sample of people born between 1980 and 1984 and first interviewed between ages 12-17 and a publicly accessible database. Our focus was on data for the 11years 2000-2010. Using whole sample data, we tested for a reciprocal association between depression and anxiety during each 2-year period and arrests during the following year, and vice versa, allowing for relatively fixed characteristics such as sex, age and socio-economic indicators. We used period-specific change modelling to test relationships. We found that within-individual increases in depression and anxiety scores over short periods (2-year periods) of time was associated with an increase in the number of arrests subsequent over the following year, consistently throughout the whole of the 10years studies. The reciprocal association was also observed, albeit the magnitude of the effects was much smaller. This study adds to the literature on the association between mental health and CJS contact by showing that they may be reciprocally associated. This suggests that facilitating co-working or even formal partnerships between community mental health services and justice-related services could be beneficial.