My guess would be that, due to increased horizontal transfer of genes (as compared to virtually all other groups of organisms), bacterial systematics cannot achieve quite the same resolution as attained elsewhere, much though there may be detailed codes of nomenclature. Specifically why those two ranks are the ones to fall out, I am less sure, but perhaps gene transfer could result in more confusion at that level than otherwise.  

In addition to the difficulty of resolution with bacterial taxa, sub-families and tribes are not always present in other groups of organisms either. For instance, some plant families have tribes (Asteraceae) or sub-families (Fabaceae), but many do not have either, and none have both.

To Serban Proches: I suspect the problem of horizontal transfer is more prevalent at a 'lower' taxonomic level such as genus (of course we need to bear in mind that ranks are artificial). If horizontal transfer is a major problem, then we should not even use the genus rank. I think it has become better supported that bacterial evolution is by and large tree-like, so it is not the lack of phylogenetic structure that has stopped people from using intermediate ranks such as subfamily and tribe. Actually these ranks were used more often in older literature, when our abilities to reconstruct phylogenies were still very limited. It is odd that now with unprecedented phylogenetic research intensity, useful ranks are not used.

To Daniel S Park: The groups you mentioned are large and intermediate ranks can be very useful. This is the problem I face with bacterial taxonomy. Some large families such as Enterobaceriaceae have no subfamilies or tribes, making it very difficult to navigate among the numerous genera. There can be several reasons as to why intermediate ranks may not be needed: the group doesn't contain many species or genera; the classification lacks revision; or different workers have different traditions of using ranks.