Roles or rules?

Jara-Ettinger and Dunham's institutional stance conceives of social behavior in terms of roles; economists instead define institutions in terms of rules. I argue that rules, rather than roles, are logically (and likely cognitively) primary because (i) rules are underdetermined by roles, (ii) many rules are more abstract than roles, and (iii) rules only work if they rarely need to be enforced.