One of the key notions entailed in the committee government model, especially in its more extreme “subgovernment” version, is that many congressional committees are unrepresentative of their parent chambers. It is easy to see why this notion is important: if most committees mirrored the range of interests found on the floor, then the autonomy that they are assumed to possess would have much fewer far-reaching consequences.

The belief that many committees are unrepresentative is partly based on a deduction from the assumptions that members seek assignment to committees pertinent to the interests of their constituents (the interest-seeking hypothesis) and that most assignment requests are routinely accommodated by each party's CC (the accommodation hypothesis). We believe the premises of this deduction to be shaky (see Chapter 2), but unrepresentative committees may arise even if members do not self-select onto committees. Self-selection focuses on the unrepresentative character of those who enter a committee, but it can also be that those who exit are unrepresentative, and either process is sufficient to produce unrepresentative panels. If we accept the possibility that subgovernments may arise whenever there are unrepresentative committees (unrepresentativeness + autonomy = subgovernment), then committee unrepresentativeness itself is of interest to those who would assess the plausibility of the view that Congress is something like a giant logroll among subgovernments.

The view that a system of reciprocity exists among autonomous committees, who trade support on the floor (or “defer” to one another's policy-area expertise), seems to require a certain number of unrepresentative committees.