« [SSJ: 263] A Banking Kabuki? | Main | [SSJ: 265] work hour reduction policy »
September 11, 1995
[SSJ: 264] RE P-A Theory and Power
From: Michael Thies
Posted Date: 1995/09/11
Hiwatari is dead right in recommending the McCubbins and Noble piece as a clear
explanation of the point of principal-agent analysis. I write to respond the the
last paragraph of his submission, in which he asks, "where do we go from
here...is the only value of P-A theory to say that politicians are in charge?"
[I use quotes, but paraphrase].
First of all,the reconceptualization of the relationship between burs and pols
as one of delegation and not as a power struggle with no indentifiable rules or
resources represents a huge step forward. Now we can ask much more precise
questions about the amount of discretion delegated, and the administrative rules
and controls that pols employ to "manage" the delegation. Ellis Krauss is
correct -- the "agents" in the bureaucracy certainly affect policy outcomes and
even the initial agenda setting. That, indeed, is the point. Politicians are not
expert in all policy areas in any political system, so they "hire" experts to
handle details and even to suggest changes. The worry of course is that agents
will use the discretion they have been delegated to go off on their own -- to
suggest, push through, and implement policies anathema to their political
"principals" and, by extension, to voters. To do so, they must first want to
(and no one has ever clearly explained why bureaucrats might want something all
that different from what politicians want -- other than maybe controlling pork
for efficiency reasons -- this is especially a problem in studies of Japanese
policy making). Then, they must have the tools to do so, and the ability to
evade political efforts to rein them back in.
The study of what tools bureacurats have been provided, and what administrative
controls politicians retain to oversee bureaucratic activities, then, is a rich
field for comparative polical studies. How, for example, do parliamentary and
presidential systems differ in terms of the discretion delegated to bureacurats
and the extent of administrative controls retained by pols? Do majoritarian
systems differ further from coalition-govt systems? Only after answering
questions like these, should we concern ourselves with explaining remaining
anomalies. For example, most of the differences between the U.S. and Japan on
these dimensions might be explained without resort to country-specific
variables. Of course, we can never explain all of the variation we see that way,
but we can go a long way toward explaining the lion's share of interesting
variation.
Other questions involve such things as variation in bureaucratic discretion and
administrative controls across issue areas WITHIN a single country. Compare tax
policy and chemical regulation in the U.S., for example.
So, I agree with Mr Hiwatari in one sense: if a study concludes that policians
are in charge BECAUSE politicians are principals and bureaucrats agents, the
study has told us nothing. Hopefully, through comparative studies , either
across issues or across countries -- or across time, using comparative
statics--we can move far beyond such useless tautology. But to reiterate the
central point of McCubbins & Noble, we must always be careful to distinguish
between activity and authority, and hence, between delegation and abdication.
Abdication is certainly possible -- just not nearly as prevalent as many have
thought, especially with regard to Japan.
Approved by ssjmod at 12:00 AM