Further Analysis of Administration

1

• Further Analysis of Administration
• Many popular perceptions of bureaucracy are
  incorrect, judgmental.
• Have been focusing on relations between
  legislature and bureaucracy.
• Now turn to behavior of bureaucrats themselves.
• Will return to theme of legislative control.

2

• Niskanen Model
• Proposed that we view a bureau as a division of a
  private firm.
• Bureaucrat is the manager who runs the division.

3

• What motivates the bureaucrat?
• Niskanen supposed that the bureaucrat seeks to
  maximize his budget.
• Why?
• Cynical view bureaucrats compensation is tied
  to size of budget.
• Also, more perks.
• Nonmaterial gratification.
• Usually, bureaucrats care about their agencys
  mission.

4

• Important point about Niskanen sinners lumped in
  with saints.
• Doesnt make a difference why bureaucrats want to
  maximize their budgets.
• Entirely non-judgmental.

5
TC

B
Q
L
Q0
Q
H
6

• Information
• Bureaucrat knows B.
• Bureaucrat knows TC.
• Legislator only knows B.

7

• Niskanen assumes that bureau must cover its
  costs--will not agree to a Q where TC gt B.
• Thus Q is in range L, H.
• Q maximizes the budget.
• Bureaucrat will choose Q.
• Q0 maximizes social benefit (with perfect
  information, legislature will choose this).

8
TC

B
Q
L
Q0
Q
H
Q
9

• Niskanen calls Q the demand-constrained level of
  Q.
• Q is the cost-constrained level.
• Argues that Q0 lt Q lt Q.
• Bureaus seek too much money.

10

• Basic point of Niskanen model
• Government is too big.
• (Q and Q gt Q0.)

11

• Model is controversial.
• Niskanen made assumptions clear.
• Many critiques of these assumptions.

12

• Motives
• Niskanen assumed budget-maximization.
• Define slack as B TC.
• If bureaucrats care only about slack then they
  will operate at Q0, but do so inefficiently.

13

• Alternatively, suppose that bureaucrats are
  interested in the quiet life.
• Self-selection process.
• Results in underproduction, or Q lt Q0.

14

• Another critique/variation on Niskanen Miller
  and Moes analysis of the active legislature.
• Niskanen assumes a passive legislature.
• Bureau requests pair (B, Q) and legislature
  goes for it.

15

• Like a car buyer telling salesman exactly how
  much she is willing to pay for each vehicle on
  the lot.
• Monopoly bureau with private information and a
  passive sponsor.

16

• Miller and Moe argued that the legislature should
  have the capacity to anticipate implementation
  problems.
• Legislators do research, too.
• Learn about bureaucracy through hearings,
  document requests, investigations by staff.

17

• Miller and Moes alternative to Niskanen
• Bureau may not have the upper hand.
• Legislature may be able to appropriate all rents
  from bureau.

18

• Principal-agent formulation of delegation game
  and rules game is a hybrid of Niskanen, and
  Miller and Moe
• Principal-agent approach combines agents ability
  to extract rents with principals legal authority
  to write the contract with the agent, delegate,
  and set rules.

19

• Bureaucratic Drift
• A new law is the work of an enacting coalition
• Legislators in the House,
• Legislators in the Senate,
• The President,
• Bureaucratic entity.

20

• Coalition has coordinated around a policy
  objective.
• Seek implementation.
• Also want durability.
• Latter is hard to obtain.

21
Example Ideal Points
B
Resources
P
S
H
Authority
22
B
Resources
P
X
X
Drift
S
H
Authority
23

• X is the policy chosen by (H, S, P).
• Points outside triangle are unanimously disliked
  by (H, S, P).
• If bureaucrat goes outside it, a new law will be
  passed or bureaucrat will be penalized in some
  way.
• For example, bureaucrat may be fired and replaced.

24

• The best the agency can do is to implement X.
• Difference between X and X is bureaucratic
  drift.
• Bureaucrat exploits
• Multiple principals.
• High transactions costs of changing law and
  monitoring.

25

• One implication of bureaucratic drift model
  changes in agency behavior are explained by
  changes in legislative preferences.
• Another implication of bureaucratic drift having
  multiple principals makes a difference.

26

• Another Look at Drift

A
Plow
P
Phigh
Within Plow, Phigh, transactions cost of
monitoring outweighs principals gain from
implementing P. Agent chooses Plow. Closest
to his ideal point.
27

• With unilateral oversight, adding another
  principal reduces agents discretionary power.

Phigh
A
Plow
P
Plow
P
Phigh
Agent acts within Plow, Phigh. Smaller
than Plow, Phigh. Agent chooses Plow.
28

• Bureaucratic Drift model assumes cooperative
  oversight.

E
L
A
Pareto Set
Adding principals can only expand Pareto Set.
With cooperative oversight, adding principals
benefits agent.
29

• Restricting Drift
• Appointment process is powerful.
• Legislators select agency heads who they believe
  to be loyal.