Within empirical approaches to racial residential segregation, there has been a tendency to draw on the work of or influenced by Robert E. Park and his ecological hypothesis to explore social and spatial relations between members of different ethnic and racial groups, thus framing research within a race relations paradigm. This has promoted an analysis which naturalizes racial differences but which also sidelines structural considerations. In turn this approach has also fed into political discourses on segregation, at times supporting more reactionary positions. This paper seeks to address this debate by considering whether emphasis on minority ethnic concentration sidelines the more pertinent issue of concentration in deprived areas, suggesting that neighbourhood deprivation as a measure can be more easily aligned with structural conditions which have influenced the settlement and historical experience of many ethnic minority communities. Specifically, I consider the extent to which a measure of neighbourhood deprivation is more important than the ethnic composition of an area for thinking about the distribution of inequalities in unemployment (as one example of socio-economic inequality). Using multilevel logistic regression I find neighbourhood income deprivation to be more important than levels of co-ethnic concentration for explaining ethnic differences in unemployment. The findings imply that neighbourhood deprivation is significantly more important for considering inequalities in unemployment for ethnic minorities than the ethnic composition of an area.