If fixed effects and robust standard errors both necessary, do they have to be at the same level, and why?

I am working on an empirical paper using repeated cross-sectional data, and a reviewer has asked that we cluster our standard errors at the same level as our geographic fixed effects. Given the structure of our data, it makes the most sense to cluster at the level of the enumeration area (based on Abadie et al, 2017 and Angrist and Pischke, 2009). But since treatment occurs at the level of the enumeration area, I believe an enumeration area fixed effect would result in multicollinearity. Originally, we used sub-district fixed effects, but this reviewer seems to believe that consequently, we need to match our standard error clustering so that it is also at the sub-district level.

I have never in my studies come across any theoretical work supporting the claim that if fixed effects and robust standard errors are both necessary, then they must be at the same level. Could anyone please point me in the direction of any relevant citations, or could anyone briefly offer an explanation as to theoretically why we want the level of the fixed effects to match the level at which we cluster the standard errors?

• Welcome to CV. Why do you think the enumeration fixed effect would result in multicollinearity? Are the subdistricts smaller than enumeration areas? Commented Sep 21, 2020 at 16:36
• Yes, the subdistricts are smaller than the enumeration areas. My estimation approach relies on a DID, and treatment is at the enumeration area level. Commented Sep 21, 2020 at 16:51
• In each cross-section, are you repeatedly drawing new samples of subdistricts? And treatment is at the enumeration level, and should affect all subdistricts embedded within enumeration areas? Am I correct? Commented Sep 21, 2020 at 17:00
• I am so sorry, I mis-wrote before. Enumeration areas are smaller than subdistricts. Each cross-section is a new sample of enumeration areas. some of these enumeration areas are treated and others are not. Apologies for adding more confusion with my typo! Commented Sep 21, 2020 at 17:09
• So the enumeration areas are nested within sub-districts. Three questions: (1) Treatment is at the enumeration area level, correct? (2) Did you use sub-district and time fixed effects in your paper? (3) If treatment is at the enumeration area level, shouldn't you be observing the same units before and after the policy/treatment? Repeated cross-sections at the lowest level won't guarantee this, unless it is 'aggregated up' to a higher level? Could you provide further clarify. I'm sure others can offer further guidance as well. Commented Sep 21, 2020 at 22:04