Designing Decision-Making Procedures in MPCs: The Advantage of Premise-Based Interest Rate Decisions
With Øistein Røisland (Norges Bank)
We study the relevance and implications for monetary policy of
a judgment aggregation problem called the 'discursive dilemma'. The
dilemma implies that the interest rate decision depends on whether the
MPC votes directly on the interest rate (conclusion-based procedure),
or reaches decisions through voting on the premises for the decision
(premise-based procedure). We find that the two decision-procedures
are likely to give non-negligible differences in policy. Normatively, our
results suggest that a premise-based procedure tends to give better
decisions unless uncertainty is strictly additive, in which case the two
procedures are normatively equivalent. MPC ownership of the infla-
tion reports and a 'core model' are institutional devices that support
premise-based policy decisions.
Communicating Monetary Policy when the MPC Members Believe in Different Stories
With Øistein Røisland (Norges Bank)
Most central banks provide the public with information about the
reasoning behind a specific interest rate decision, that is, a story.
When decisions are made by a committee, it can be problematic to
find a story that is both consistent with the decision and representa-
tive for the committee. The paper goes through different alternatives
of finding a consistent story. We show that the only alternatives that
give unique and consistent stories are: (i) vote on the interest rate and
let the winner decide the story, (ii) vote on the elements of the story
and let the interest rate follow from the story. The two procedures
tend to give different interest rate decisions and different stories, due
to an aggregation inconsistency known as the "discursive dilemma".
We investigate the quality of the stories under the two approaches,
and find that alternative (ii) gives stories that tend to be closer to
the true, but unobservable, story. Thus, our results give an argument
in favour of premise-based, as opposed to conclusion-based, decision-
making. The paper also discusses institutional devices that support
premise-based decision-making, such as a core forecasting model and
an inflation report "owned" by the MPC.
A Quantitative Discursive Dilemma
with Øistein Røisland (Norges Bank). (Revised version of Norges Bank WP 7/2007. pdf
The typical judgment aggregation problem in economics and other fields is the following: A group of people has to judge (estimate) the value of an uncertain variable y which is a function of k other variables, i.e. y = D(x1,...,xk) . We analyze when it is possible for the group to arrive at collective judgments on the variables that respect D. We consider aggregators that fulfill Arrow's IIA-condition and neutrality. We show how possibility and impossibility depend on the functional form of D, and generalize Pettit's (2001) binary discursive dilemma.
Collective Economic Decisions and the Discursive Paradox.
With Øistein Røisland, Norges Bank Working paper no. 3/2005 (pdf)
Most economic decisions involve judgments. When decisions are
taken collectively, various judgment aggregation problems may oc-
cur. Here we consider an aggregation problem called the discursive
dilemma, which is characterized by an inconsistency between the ag-
gregate judgment on the premises for a conclusion and the aggregate
judgment on the conclusion itself. It thus matter for the decision
whether the group uses a premise- or a conclusion-based decision-
making procedure. The current literature, primarily within jurispru-
dence, philosophy, and social choice, consider aggregation of qualita-
tive judgments on propositions. Most economic decisions, however,
involve quantitative judgments on economic variables. We develop
a framework that is suitable for analyzing the relevance of the dis-
cursive dilemma for economic decisions. Assuming that decisions are
reached either through majority voting or by averaging, we
nd that
the dilemma cannot be ruled out, except under some restrictive as-
sumptions about the relationship between the premise-variables and
the conclusion.