Political Control of the Bureaucracy I

1
Political Control of the Bureaucracy I

• David Epstein
• EITM Michigan
• June 18-19, 2003

2
The Problem Base Closing

• Near the end of Cold War Government wanted to
  close some military bases
• Congress had blocked similar action after the
  Vietnam War
• Required full EITM statements for any closings
• Even then, hard to avoid Congress retaliating
• But now savings were big ? 2 to 5 billion per
  year
• Sentiment was growing for some action

3
The Solution BRAC Attack

• Congress responded by creating the Base
  Realignment and Closing Comm.
• Independent Regulatory Agency
• 12 members, chosen by Secy of Defense
• Recommended closings and realignments
• Secy of Defense had to accept or reject entire
  list without amendments
• If accepted, Congress had 45 working days to pass
  a joint disapproval resolution
• Subject to veto, with no amendments possible

4
The Puzzle Why this solution?

• Congress could have
• Ignored the problem
• Passed legislation by itself
• Delegated to the Secretary of Defense directly,
  rather than the commission
• Why did it choose this particular form of
  delegation, and to what end?
• Note Craig and BRAC-proofing in Idaho

5
Objectives

• Walk through the genesis and execution of our
  book, Delegating Powers
• Background literatures and early theory/tests of
  delegation
• Discovering the idea of the book
• Executing theory/tests/implications
• Also look at how literatures develop as
  conversations on particular topics
• Research inserting yourself into stream

6
Game Plan

• Literature(s) Review of Policy Making
• Legislative Organization
• Delegation and Constraints
• Examples of theories tests
• DP Theory
• Background Transaction Cost Economics
• Review model in detail
• Derive predictions
• DP Empirics
• Look at tests of hypotheses (homework)

7
Legislative Org. Literature
8
Legislative Org. Literature
SustainingLogrolls
9
Legislative Org. Literature
SustainingLogrolls
Committees
10
Legislative Org. Literature
SustainingLogrolls
Committees
UniversalLogrolls
Committees
11
Legislative Org. Literature
SustainingLogrolls
Committees
UniversalLogrolls
Committees
AsymmetricInformation
12
Legislative Org. Literature
SustainingLogrolls
Committees
UniversalLogrolls
Committees
AsymmetricInformation
Committees
13
Legislative Org. Literature
SustainingLogrolls
Committees
UniversalLogrolls
Committees
AsymmetricInformation
Committees
Majoritarian
Median Voter
14
Legislative Org. Literature
SustainingLogrolls
Committees
UniversalLogrolls
Committees
AsymmetricInformation
Committees
Majoritarian
Median Voter
Passing theParty Agenda
15
Legislative Org. Literature
SustainingLogrolls
Committees
UniversalLogrolls
Committees
AsymmetricInformation
Committees
Majoritarian
Median Voter
Passing theParty Agenda
Committees
16
Legislative Org. Literature
SustainingLogrolls
Committees
UniversalLogrolls
Committees
AsymmetricInformation
Committees
Majoritarian
Median Voter
Passing theParty Agenda
Committees
PartisanLogrolls
Majority Party Caucus
17
Legislative Org. Literature

• Theories disagree on fundamental problem

18
Legislative Org. Literature

• Theories disagree on fundamental problem
• All agree that committees are the solution

19
Limits of Existing Theories

• These may all capture elements of reality
• But as stated they must be incomplete
• Legislators in all countries face similar
  problems in getting reelected
• Yet not all have favored committees at the
  expense of parties
• Therefore, our system of strong committees must
  derive from broader governmental system,
  including SOP

20
Delegation Literature Two Major Strands

• Why delegate?
• Save time/Reduce workload
• Take advantage of agency expertise
• Protect special interests
• Shift the blame
• Who controls delegated authority?
• Transmission Belt Theory (legal)
• Bureaucracy (abdication)
• Congress (congressional dominance)

21
Bureaucratic Drift

• Control drift by
• Ex-ante Controls
• Deck Stacking
• Autopilot
• Ongoing Controls
• Fire Alarms
• Police Patrols

22
Delegation Theory Testing

• Move to formalize these theories in early 1990s
• Lupia McCubbins
• Bawn
• Epstein OHalloran, etc.
• Led to
• Richer, more nuanced views of delegation
• Deeper insight into particular policy areas
• Revised views of related literatures

23
Example I Delegation Trade Policy

• Analysis of trade policy had been mainly
  economic, international in focus
• Economic gains from free trade one of the central
  tenets of economic theory
• International institution (GATT) as vehicle for
  gradual lowering of worldwide tariffs
• Led by US in post-war era
• But trade authority in US is delegated
• And terms of delegation have changed
• Implies a more political, domestic angle

24
Trade Policy History

• Congress has ceded the executive substantial
  authority in trade policy
• 1934 Reciprocal Trade Agreement Act (RTAA)
• 1962 Trade Expansion Act
• 1974 Trade Reform Act
• Terms of delegation revised periodically
• Sometime Congress delegates broad authority
• E.g., RTAA gave bilateral negotiating powers
• Sometimes Congress limits this authority
• E.g., fast track procedures require Congress to
  vote

25
Basic Delegation Model

• Actors Congress (C) President (P)
• Preferences ui(x) -(x-xi)2 on XR1
• xC0, xP0
• Outcomes xp ?, ?U-R,R
• So role for executive expertise
• Order of Play
• Congress sets policy pC, discretion d
• President observes ?
• President chooses p s.t. p- pC ? d

26
Basic Delegation Model

• In equilibrium
• President moves policy as close to xP as
  possible, given d
• Congress sets pC0
• Congress sets dR-xP if xP

d
d
xP
xC
R
-R
27
From Theory to Test

• Theory Divided government leads to less
  delegation.
• President has less protectionist preferences than
  median member of Congress
• So less delegation leads to higher tariffs

28
Major Postwar Trade Acts
Major U.S. Trade Legislation, Divided Government,
Delegation, 1948-1992
29
Delegation and Tariffs

• Want to estimate a model like
• Tarifft a b1Delegt b2Econt
• But we cant measure delegation directly
• Can only measure change from year to year
• So instead, estimate
• DTarifft a b1DDelegt b2DEcont
• Same coefficients, so tests are the same
• Constant is now trend over time

30
Effect of Delegation on Tariffs
31
Example II Delegation, Fire Alarms Committees

• Fire Alarms/Police Patrols proved to be a useful
  prism for analysis
• But its clearly not an equilibrium
• If every fire alarm were answered, then there
  would be lots of false alarms
• So, do a strategic version of the model
• Has spillover implications for committees
• As we saw, literature flunks comparative test
• So predict degree of committee outliers

32
Strategic Oversight Model

• 3 Players Floor, Committee, Agency
• Ui(x) -(x-xi)2 , ?i?F,C,A
• xF 0, xA 0
• Incomplete information x r ?
• ? U-1,1, status quo x0 0
• Order of play
• Agency proposes regulation (r)
• Committee recommends accept or reject (m)
• Floor overturns policy or lets it stand (?)

33
Strategic Oversight Model

• 3 Players Floor, Committee, Agency
• Ui(x) -(x-xi)2 , ?i?F,C,A
• xF 0, xA 0
• Incomplete information x r ?
• ? U-1,1, status quo x0 0
• Order of play
• Agency proposes regulation (r)
• Committee recommends accept or reject (m)
• Floor overturns policy or lets it stand (?)

Know ?
34
Strategic Oversight Model

• 3 Players Floor, Committee, Agency
• Ui(x) -(x-xi)2 , ?i?F,C,A
• xF 0, xA 0
• Incomplete information x r ?
• ? U-1,1, status quo x0 0
• Order of play
• Agency proposes regulation (r)
• Committee recommends accept or reject (m)
• Floor overturns policy or lets it stand (?)

Know ?
Sees r, m
35
Equilibrium Properties

• Equilibrium depends on the value of xC
• xC xA
• No info transmission due to collusion
• xF
• Noisy signaling ranges, like Crawford-Sobel
• xC
• r and m together perfectly imply ? (F informed)
• Agency modifies its proposal to ensure
  Committees support (just like fire alarms)
• Floor prefers xC - xA (so still some slippage)

36
Committees as Contrary Outliers
Ideal Point
A
C F
Like Romer-Rosenthal when CF
Status Quo
37
Committees as Contrary Outliers
Ideal Point
A
C F
Accommodation
Like Romer-Rosenthal when CF
Status Quo
38
Committees as Contrary Outliers
Ideal Point
A
C F
Accommodation
Status Quo
Like Romer-Rosenthal when CF
Status Quo
39
Committees as Contrary Outliers
Ideal Point
Agency Ideal Point
A
C F
Accommodation
Status Quo
Like Romer-Rosenthal when CF
Status Quo
40
Committees as Contrary Outliers
Ideal Point
A
C F
Now change the committees ideal point
Status Quo
41
Committees as Contrary Outliers
Ideal Point
A
F
C
to be slightly contrary to the agency.
Status Quo
42
Committees as Contrary Outliers
Ideal Point
A
F
C
This lengthens the accommodation range
Status Quo
43
Committees as Contrary Outliers
Ideal Point
A
F
C
and the status quo range.
Status Quo
44
Committees as Contrary Outliers
Ideal Point
A
F
C
Overall, outcomes that were formerly at A
Status Quo
45
Committees as Contrary Outliers
Ideal Point
A
F
C
Overall, outcomes that were formerly at A
Status Quo
46
Committees as Contrary Outliers
Ideal Point
A
F
C
are now closer to F, making the floor better
off.
Status Quo
47
Committees as Contrary Outliers
Ideal Point
A
F
C
This process continues until C -A.
Status Quo
48
Committees as Contrary Outliers
Ideal Point
A
F
C
Now neither deviation improves the floors
utility.
Status Quo
49
Committees as Contrary Outliers
Ideal Point
A
F
C
So the floor is best off when C-A. This leads to
Status Quo
50
Committees as Contrary Outliers
Ideal Point
A
F
C
So the floor is best off when C-A. This leads to
Status Quo

• Committees will move contrary to changes in
  executive preferences

51
Committees as Contrary Outliers
Ideal Point
A
F
C
Note that the range of status quos increases with
C.
Status Quo
52
Committees as Contrary Outliers
Ideal Point
A
F
C
If hearings are held for new policy initiatives,
then
Status Quo
53
Committees as Contrary Outliers
Ideal Point
A
F
C
If hearings are held for new policy initiatives,
then
Status Quo

• Outlying committees hold fewer oversight hearings

54
Testing the Hypotheses

• We can measure committee and agency outliers
  relative to the floor

• Or relative to the majority party

55
Committee Outliers Measures
Distribution of House Preferences
Distribution of Committee Preferences
Committee Median
Majority Party Median
Majority Party Contingent Median
House Median
Committee-Floor Outlier
Committee-Party Outlier
56
Testing the Hypotheses

• We can measure committee and agency outliers
  relative to the floor

• Or relative to the majority party

• Either way, the first term

57
Testing the Hypotheses

• We can measure committee and agency outliers
  relative to the floor

• Or relative to the majority party

• Either way, the first term

58
Testing the Hypotheses

• We can measure committee and agency outliers
  relative to the floor

• Or relative to the majority party

• Either way, the first term should be negatively
  related to the second term

59
Testing the Hypotheses

• We can measure committee and agency outliers
  relative to the floor

• Or relative to the majority party

• Either way, the first term should be negatively
  related to the second term

60
Data

• All House committees, 80th through 102nd
  Congresses
• Use Nominate scores for ideal points
• President scored according to positions taken on
  roll call votes
• Hearings data from Congressional Information
  Service Masterfile
• 80th91st review in content descriptor
• 92nd102nd oversight (all lower case)

61
Descriptive Statistics
62
Predicting Contrary Outliers
63
Predicting Contrary Outliers
64
Predicting Contrary Outliers
65
Predicting Contrary Outliers (continued)
66
Post-Estimation Checks

• Divided data into two samples
• President is a liberal outlier
• President is a conservative outlier
• Larger effect with conservative presidents
• Can reject competing hypotheses
• that committee medians floor medians
• that committee party medians floor party
  medians
• But differences are less than distributive theory
  would predict (