Understanding a country’s past, present and expected future population diversity at sub-national levels is important. Studying changing diversity in terms of the groups that constitute a population and how that varies within regions and between local areas assists in understanding socio-economic and demographic sub-national trends. It is important to project the probable future population diversity in regions for successful policy planning and implementation, group-specific investments in health, education, and community services, as well as the provision of non-government services. This thesis is a compilation of four inter-related studies that examine the ethnic makeup of the Auckland region of New Zealand. Using New Zealand Census of Populations and Dwellings data (1991, 1996, 2001, 2006 and 2013), the first study identifies and empirically demonstrates the shortcomings in traditional measures commonly used for measuring residential sorting, and instead proposes an alternative preferred measure. Specifically, the study shows that the Entropy Index of Systematic Segregation is the measure of residential sorting that is least biased by group size. Using the same data, the second study examines the long-term patterns in ethnic and economic residential sorting in Auckland at a fine geographic scale using disaggregated groups. The results show that residential sorting by ethnicity is much more prominent than sorting by economic factors. The results also show that, although residential sorting has been declining over time in Auckland, specific ethnic groups like the Chinese and Indians have become more residentially sorted over time. The New Zealand European, Other European, and New Zealand Māori groups were found to be the least residentially sorted, whereas small ethnic groups like the African, Latin American/Hispanic, Tokelauan, and ‘Other Pacific Island’ groups were the most residentially sorted, over the whole study period. The results also show that the dominant feature of residential sorting in Auckland is the sorting of subgroups (e.g. Chinese, Indian and South East Asian) within broad ethnic groups (e.g. Asian). Using the New Zealand Linked Census (NZLC) data, the third study sheds light on the factors that predict the self-identified ethnic affiliation of adolescents in Auckland. The results show statistically significant relationships between the adolescents’ ethnic identity and the ethnic identity assigned to them five to seven years previously by their parents. Additionally, the ethnic affiliation of adolescents is also associated with their age, sex, having been born in New Zealand, the ethnic makeup of the neighbourhood they live in, and their parents’ ethnic identities. The results confirm patterns of complementarity between ethnicities and ethnic groups that are consistent with other research. Finally, using NZLC data, the fourth study in this thesis describes the construction and calibration of a spatial microsimulation model, which can be used to project the expected future ethnic residential sorting and ethnic diversity in Auckland. Results show that our model is capable of reproducing the dynamics of residential sorting in Auckland with minimum error.