No.
They may have held back on the memory clock because they didn't need more bandwidth for this part and to have this increased memory bandwidth as a selling point (oh, it looks so pretty on the box/spec sheet) for the R580, but that's it.
If you look at the rumored specifications for R580 and comapre them to R520 you will see that R580 (like RV530) marks a special shift in the performance mix with hree independent shaders per pipe as opposed to a single one. This means that R580 should be a shading beast compared to R520 and that no clocks on the R520 could save it from utter defeat by its sibling. What this means, in my opinion, is that the R520 represents a new implemenatation of the classic configuration and proportions, while the R580 will be part of the new wave of incredibly shading capable GPUs. Because of this, no clocks on the R520 really matter as this will mark a changing of the guard.
I do think that ATI may have used slightly more conservative numbers on the R520 than they initially planned, but I don't think R580 is the reason. Instead, I believe the reason to be that R520 has been a "problem child" from the start and there were no incentives to take more risks of it becoming one in retail. Lower clocks mean higher yields and fewer problems with the products sold. Both of these factors help profits as long as the part is still competitive, which it is. This is supported by thesupposed last-minute change from a 650MHz core clock to 625MHz, a small degradation in performance that probably makes yields, power consumption, heat, and RMAs more manageable.
