Shipan: Regulatory regimes, agency actions and the conditional nature of political influence
Disclaimer. Don't rely on these old notes in lieu of reading the literature, but they can jog your memory. As a grad student long ago, my peers and I collaborated to write and exchange summaries of political science research. I posted them to a wiki-style website. "Wikisum" is now dead but archived here. I cannot vouch for these notes' accuracy, nor can I even say who wrote them. If you have more recent summaries to add to this collection, send them my way I guess. Sorry for the ads; they cover the costs of keeping this online.
Shipan. 2004. Regulatory regimes, agency actions and the conditional nature of political influence. APRS.
The literature has found conflicting results as to whether agencies are responsive to the current Congress. (Current Congress--so we're not worried about APAs and agency structure here.) Shipan uses a spatial model to show that reponsiveness depends on the alignment of preferences ("regime") among the agency, committee, and Congress.
- Agency (A)
- Committee (C)
- Floor (F)
Order of Play
- Agency proposes an action, 'a'
- Committee either "gatekeeps" (takes no action) or passes a bill.
- If committee passes a bill, then floor hears it with an open rule and amends it to match its ideal point.
- Possible outcomes:
- Agency passes 'a', committee gatekeeps (does nothing), end.
- Agency passes 'a', committee passes a bill, floor modifies bill to equal F, end.
- Unidimensional space:
- F* . . . . . . C . . . . . . F
- Since F* is as far from C as F is, C (committee median) is indifferent between F (floor median) and F*.
- In the article, Shipan denotes F* as C(F).
- Preferences in this order:
- A . . . . C(F) . . . . C . . . . F
- If A proposes 'a' < C(F), the committee will pass a bill. The floor will amend it. Result: F.
- If A proposes 'a' = C(F), the committee is indifferent between 'a' and F, so it has no incentive to pass a bill. Result C(F).
- THUS: A passes 'a' = C(F). It is responsive to the committee.
- As C gets further from F, the range of C(F) to F gets larger. Thus, a larger gap between C and F makes the agency less responsive to F.
- THUS: Agency is somewhat responsive to Congress.
- Preferences: A is between C(F) and F. Example:
- C(F) . . A . . C . . . . . F
- Result: A can always propose its ideal point, and C never has an incentive to act. Thus, A always gets what it wants.
- THUS: Agency is not responsive to Congress.
- C(F) . . . . . C . . . . . F . . . . . . . . . . . . . A
- Result: A proposes its ideal point. C will always pass a bill. F will always amend the bill to match its ideal point. Result is always F.
- THUS: Agency is perfectly responsive to Congress.
- Shipan ignores presidential veto power. It's too complicated for this model.
- President can influence A's ideal point (through appointments, reorganization, etc.).
- Thus, although A does not necessarily equal president's ideal point, A will have the same relationship to C and F that P has.
- Thus, under the GATEKEEPING regime, the president's ideal point will predict agency outcomes.
Doesn't really change anything.
Examines the FDA over several decades.
- Y = Number of FDA investigations. More investigations means the FDA is more liberal (huh? Why? Is this valid?).
- Using ideology scores for presidents, committee means, and floor means, Shipan codes each year as belonging to one of his three regimes.
- Interactive model. See table 3.
- Under the committee-floor regime, the committee has a positive influence and the floor has a negative one, as predicted.
- Under the gatekeeping regime, the president has a weakly positive influence; in the right direction, but not as strong as predicted.
- Under the floor regime, the floor has a strongly positive influence.
- The theory-based regression (Table 3) doesn't seem to fit the data better than the atheoretical regression (Table 2; compare R2). This makes me think that it's the control variables that are explaining all the variance in Y, not the main independent variables.
- Is the operationalization of Y really valid? There is no justification offered that we should think so.