Bawn: Political control versus expertise

Bawn. 1995. Political control versus expertise: Congressional choices about administrative procedures. APSR 89:62-73.

MAIN POINT

When delegating to agencies, Congressional coalitions (C) must minimize either procedural uncertainty (what will the agency do?) or technical uncertainty (is this the best regulation?); there is a tradeoff. Thus, the degree of agency independence (i.e. the strictness of of procedural rules) (Y) "reflects the legislature's willingness to trade uncertainty about policy consequences for uncertainty about agency behavior" (p 63).

PROCEDURAL (POLITICAL) UNCERTAINTY

Administrative procedures determine which external "signals" an agency hears: does it listen mostly to the industries it regulates (as in "capture" theory), to oversight committees (congressional dominance theory), or to itself (bureaucrat's ideology or professionalism)? Because "procedures determine the relative strenghts of signals received from different groups" (p 65), procedures can determine "the location of an agency's ideal point." However, there is uncertainty for C in how procedure's will affect A's (the agency's) ideal point. Thus, C uses procedures to determine the distribution of ideal points (both its mean and spread) from which A's ideal point will be drawn.

TECHNICAL UNCERTAINTY

Increasing independence allows A to incorporate more knowledge of policy consequences into its decisions. Thus, technical uncertainty falls as political uncertainty (independence) rises.

LIMITATIONS/CONCERNS

• The model treats C as a single decision maker with coherent preferences. Let's suppose you relax this assumption and let it vary. When C's unity/coherence is low, the model will be less accurate in its predictions, right? Wouldn't the variation in C's unity covary with one of the variables in the model (political uncertainty)? If so, the model's accuracy co-varies with one of the model's variables (heteroskedasticity). Hmm.
• Is there really a tradeoff between political uncertainty and technical uncertainty? Can't procedural rules be used to improve (technical) policy consequences (not only to preserve political control)? Procedural controls can force A to conduct more/better research before changing a policy. Fire alarms can be used to promote policy quality, not just political control.