Gender expression in Coprosma is not as firmly fixed as people think. Indeed, Allan (1961, FL I) even states this. I have seen solitary Coprosma repens with male and female flowers (indeed there is an example of this loaded up on the NZPCN website), or this species with seemingly 'perfect' hermaphroditic flowers. Others have reported this in C. foetidissima, C. rubra, C. spathulata subsp. spathulata, and also, as you observe in C. robusta (which I have seen too). In Coprosma repens I have noted that in some examples the anthers of 'perfect' flowers are not functional, or on plants with both genders the 'male flowers' lack functional stamens (but not always either). All of this would be worth studying further, especially as I also believe many of our Coprosma have some degree of apomixis as well (C. robusta especially so). Peter Heenan has shown this for C. waima and knows of other examples.