The article has a very suggestive title; deconstructing power asymmetries. The second part of the title suggests that by understanding the perceived level of influence one can know more about how power asymmetries operate. However, the article does not go very far in this regard. For once, a central element in the EGF is the connection of the multiple structures of governance with real and perceived policy outcomes \cite{Lubell_2013}. If we agree that water management committees are created to deal with problems that require collective action, then it will be mainly in the measurement of positive outcomes for those with relatively less power, that we will find strong indications that participation reduces power asymmetries. Additionally, if researchers cannot discard that the high perceived levels of influence of less powerful actors are produced by the strategies powerful actors use to select the composition of committees and the roles assigned in these committees, then there seems to be only weak evidence that less powerful actors can change existing asymmetries. On page 24, authors argue that elite capture is not likely because the perceptions of influence that actors have of others, does not necessarily favor governments and organized users with resources. The fact that the forums under study offer opportunities for newcomers to become influential, or that actors who have spent the most time in the committee are not the most influential, does not necessarily discard elite capture. Elites can still filter who enters the committee, they could even assign roles, and still weaker actors could develop positive perceptions of their influence. This is somewhat confirmed by the fact that some powerful actors can overrule committees decisions if they deem necessary during a crisis (See pg. 23). In conclusion, authors should develop a stronger counter-argument against the elite capture hypothesis.
In the methods section, authors discuss the sample used in the analysis. Although 93% response rate sounds great, readers don't know the total number of people involved in these forums. If 45 people were included in the sample, how many people participate in the forums in total? Additionally, I recommend introducing a note about who was left out of this sample and why. It will be of particular interest for readers to know what kind of actors excluded themselves from the study and to know more about the profile of those who have not attended at least two of the last six plenary meetings. Were they powerful actors or not? If they were not, could the study be leaving those with less power underrepresented? The same section is missing a specification about the number of coders that intervened in the thematic analysis. How were issues of intercoder reliability handled? This is of particular importance since interviews were likely conducted in Portuguese. Were translations of interviews used at all?
The case under study shows some very particular characteristics, such as the size of the cities served by the watershed, the relative power of the state governments, the years of operation in the committee. However, the discussion and conclusion sections do not tackle issues of generalizability of results to other contexts in the global South.
When authors contrast their findings with those of Mewhirter et al., authors emphasize the importance of considering the context where water partnerships operate. However, the characterization of the context where the research took place lacks some relevent information. Can we assume that data collection occurred in a typical year for these institutions? Was this context free of major environmental, political, or economic disturbances that could modify the interpretation of the results?