Huber: Delegation to civil servants in parliamentary democracies

Huber. 2000. Delegation to civil servants in parliamentary democracies. European Journal of Political Research 37:397-413.

In Brief

The principal-agent (p-a) model can be used to structure a research agenda on delegation to civil servants in parliamentary democracies. Generally speaking, the p-a model would predict that variations in political and institutional factors (X) would affect the means of delegation that ministers employ (Y). This article doesn't try to prove anything so much as to show how to apply the p-a model to parliamentary democracies.

For example: In situations of high portfolio instability (X), ministers may have trouble controlling their bureaucrats. Thus, global budgets (Y), an ex ante control device, become more common.

Principals and Agents

The principal-agent framework considers ex ante and ex post mechanisms used to keep agents acting in accordance with principals' beliefs.

• Ex ante:
  • Give ministers control over personnel; "Politicize" the civil service
  • Administrative procedures, statutes, executive orders, ministerial decrees, etc.
• Ex post:
  • Police patrols and fire alarms, etc.

Budgets can work both ex ante and ex post

Principals and Agents in Parliamentary Demoracies

Ministers don't have total control over their ministry. There are other actors (courts, other ministers, the budget process). This can be brought into the p-a framework with the notion of multiple principals.

Cabinets may be one-party or coalition governments, which could also play a role--especially for actions that require cooperation between two or more ministers (if they are from different parties). Again, multiple principles.

Two Implications of P-A

1. "We cannot draw precise conclusion[s] about how specific features of the political and institutional context affect delegation outcomes without a clear understanding of how ministers in the cabinet respond to these features of the political and institutional context when delegating" (p 404).
   • In other words: We can't study how a political/institutional factor (e.g. cabinet instability) affects a minister's control of a ministry... unless we first figure out that the minister has anticipated the problems caused by cabinet instability and done something about them. You can't just say, "How does cabinet instability affect control of the ministry?" You must say, "What do ministers do to control their ministries despite cabinet instability?" We would expect ministers do have chosen delegation structures that limit the "agency cost" of cabinet instability.
2. The p-a framework "suggests an alternative avenue for testing principal-agent theories in the cabinet context" (p 404).
   • Previous models (Weber's, historical institutionalism, organizational theory, etc) predict either no variation or random variation between the political environment (X) and delegation institutions (Y). The p-a framework does.
   • For example: In situations of high portfolio instability (X), ministers may have trouble controlling their bureaucrats. Thus, global budgets (Y), an ex ante control device, become more common.