In the model, team morale and stakeholder satisfaction are actually intended as indicators of team effectiveness. This is based on a commonly-used definition of team effectiveness in team research that was ventured by Hackman in 1978. They are like the surface symptoms of an effective team, which in turn is “caused” by the preceding factors in the model. Think of them as surface symptoms of an effective team.

I fully agree that it probably goes both ways. Teams with high morale are more effective, and high morale makes teams more effective. Ultimately we have to go with a model that we think best fits existing definitions and theories, and that's what we did here. “Job satisfaction”, for example, is also usually defined as an outcome of team processes.

The problem with causal effects is that they are in practice extremely hard to “prove”. It requires longitudinal tracking of teams and ruling out all other factors that are not in the model. That is nigh impossible. In practice, I think it's more helpful to think of these concepts as highly correlated. That is; an increase in one will likely see an increase in the other. That isn’t causal necessarily, but in practice, it results in mostly the same interventions. Because if you want to increase team morale through team effectiveness, you’ll probably end up with similar kinds of things as when you want to increase team effectiveness through team morale. And if not, that makes for a nice experiment

Sorry, long answer. But I hope it helps :)