Quantitative study of the permanent sample plot (PSP) database can provide insights into growth, mortality and recruitment processes driving forest dynamics. Modelling the dynamics of forest growth and yield provides opportunities for optimising silvicultural systems and generating accurate growth and yield estimates, which are fundamental to sustainable forest management. This paper will outline model development based on analysis of a large native forest permanent sample plot database in Papua New Guinea. We quantify the competitive influences affecting individual tree growth and mortality, and build predictive models for growth and mortality based on a hierarchical Bayesian modelling methodology. This method allows the parameterisation of a global model with species-specific parameters; therefore, species-level growth and mortality traits are preserved in model predictions, even for rare species. We examine a range of spatial and non-spatial competition indexes for the data, and conclude that a simple non-spatial competition index (basal area of competing trees within 20 metres of the subject) adequately characterises competitive influences on growth and mortality. In future work, species-specific model parameters can be used as the basis of a forest simulation system (see http://twoe.org for developments) to improve the design and intensity of selectiveharvesting regimes at the community and the concession level.