Observations from networks of forest plots show important differences in biomass and turnover of forests across continents, and how these dynamics are changing through time. These patterns are not well captured by the current generation of Earth System Models (ESMs). Vegetation demographic models (VDMs) improve the representation of vegetation within ESMs, for instance by accounting for vertical competition for light, and successional dynamics of plant functional types (PFTs). However, benchmarking competitive dynamics between PFTs presents a number of challenges, especially in species rich tropical forests, where trait data may not be available to inform the assignment of species to PFTs. We present a framework for benchmarking demographic rates and forest structure in the VDM FATES, the Functionally Assembled Terrestrial Ecosystem Simulator, using observations from forest plots. At the site level, we assign species to early and late successional strategies based on demographic rates, and compare observations of growth, survival, aboveground biomass, and size-distributions with two PFT FATES simulations. We then compare pan tropical FATES simulations to observations of aboveground biomass and woody biomass turnover rates and discuss possible causes of model biases. Forests exchange vast amounts of energy, water and carbon with the atmosphere, and in doing so play a critical role in regulating the global climate. To predict future forest dynamics, and their impacts on climate, VDMs must be able to capture trends in forest dynamics driven by changing demographic rates.