Many actions don't have a clear party (parent, teacher, boss, etc.) waiting to reward or punish, though, and the research is specifically about self control. While some people will take advantage of others' responsible behavior, self-sabotaging behavior probably has a greater overall impact on most peoples' lives.

Good example, by the way. I hadn't picked up on it reading the article, but the study really does assume an immediate observer is going to be rewarding the child, rather than more abstract forces such as the job market. The latter is hard to duplicate in a self-contained experiment, of course.