I don't think his measure of difference is comprehensive:

• The higher chance of finding smart collaborators increases chances of increased productivity
• A larger chance of making very significant improvement (a highly competitive field is probably much closer to a field-wide, world changing epiphany - while in a less competitive field, much time must be wasted laying down the groundwork)
• A longer productive life-span (much likelier to find smart assistants/students to teach at maximum ability all life long)
• A higher utility to society - the field is likely competitive because of large public attention, which in turns signals large groups of people funding research, in turn showing that smaller improvements are considered much more valuable than in other fields
• A wider selection of interesting work. It's much more likely that relatively minor or mundante results/problems in the competitive field are going to be immediately useful/used