Inefficient voting with identical voters
| Author | Ole-Andreas Elvik Naess |
| DOI | 10.1177/09516298221085973 |
| Published date | 01 April 2022 |
| Date | 01 April 2022 |
| Subject Matter | Articles |
Inefficient voting with identical
voters
Ole-Andreas Elvik Naess
Centre for Applied Research at Norwegian School of Economics
Abstract
This paper constructs a two-period electoral model where the future is uncertain and the policy in
each period is decided by majority rule. The voters’optimal future policy takes into account all
possible future realizations, while the future pivotal median voter only cares about the median
value of the future distribution. This dynamic conflict of interest implies that the electoral out-
come may be strictly Pareto-dominated by other policies even in cases where all voters are iden-
tical and have the same beliefs for the future. I apply the model to analyze the dynamic incentives
for redistribution.
1. Introduction
Consider a society where all voters are identical and have the same rational beliefs for an
uncertain future. Although all voters are equal and the set of voters is unchanged in the
future, I show that there may still exist a dynamic conflict of interest in this society. The
voters today care about all possible future outcomes, but only the pivotal median voter
will influence the future electoral decision. This means that a dynamic conflict of interest
arises between the entire future distribution and the median value of the same distribution.
I construct a two-period model, where individual future wealth is stochastic, and the
electoral outcome is decided by majority rule in each period. In each period, there is an
election where the voters choose whether to make an investment. All voters are identical
and have the same beliefs for the future, but I still show that the equilibrium policy choice
may be strictly Pareto-dominated. All voters today know that the future median voter,
Corresponding author:
Ole-Andreas Elvik Naess, Centre for Applied Research at Norwegian School of Economics.
Email: Ole-Andreas.Naess@snf.no
Article
Journal of Theoretical Politics
2022, Vol. 34(2) 175–190
© The Author(s) 2022
Article reuse guidelines:
sagepub.com/journals-permissions
DOI: 10.1177/ 09516298221085973
journals.sagepub.com/home/jtp
which is not representative for the entire distribution of their future outcomes, will be
pivotal. All voters rationally infer that the future policy choice is given by maximizing
the payoff for the particular, median realization of the future distribution. Today, the
voters choose the optimal investment policy given that the future investment is
decided by such a pivotal median voter.
I extend the model by analyzing the preferences for redistribution when I addition-
ally allow for initial wealth differences between the voters. The above argument states
that the current median voter expects to have interests that differ from the future
median voter. Typically, the current median wealth (or income) also differs from
the current average wealth (or income). Hence, two conflicts of interest occur.
There is a conflict between the current and future median voter and another conflict
between the current median voter and the current average voter. The equilibrium
outcome depends on the trade-off between these two effects. This reasoning
implies that inequality consists of two distinct concepts in dynamic settings. Future
wealth inequality is given by the ratio between the current median and average
wealth, multiplied by the future increase in inequality. Only the first term is relevant
for the current median voter, while both terms are relevant for the future median voter,
which means that the concept of inequality has different meanings for the current and
future median voter.
The current median voter knows that the future median voter has stronger
preferences for redistribution than the current median voter expects to have in the
future. I introduce commitment to the redistribution model, and I show that the
current median voter may prefer commitment to avoid redistribution in the future.
When commitment is costly, the equilibrium outcome of committing to avoid redistri-
bution is strictly Pareto-dominated by the outcome where the future median voter
chooses to avoid redistribution. However, such a promise is not credible once the
future election is reached, which means that the current median voter prefers costly
commitment.
The underlying intuition is that in the current time period, the median voter is in some
sense representative of the population (exactly half of the population are richer and poorer
than the median voter), but the preferences of the future median voter do not have such
attractive features. Today, the median value of the future distribution is just one random,
possible outcome that happens to be pivotal under the particular institution given by
majority rule.
One central motivation for this paper is to understand if a dynamic median voter model
can be used to understand the apparently puzzling relationship between inequality and
redistribution. There has been a large increase in inequality over the last decades, particu-
larly in the US (Piketty and Saez, 2003; Saez and Zucman, 2019). More economic
inequality should lead to more redistribution when there is a pivotal median voter
(Black, 1948; Downs, 1957; Meltzer and Richard, 1981). However, in many countries,
higher inequality has been accompanied by constant or decreasing redistribution (De
Mello and Tiongson, 2006; Bonica et al., 2013). This paper tries to cast some light on
this puzzle by highlighting the changing nature of the median voter in a dynamic
model. More broadly, the research question is to understand how inefficient outcomes
may arise under majority rule in dynamic settings.
176 Journal of Theoretical Politics 34(2)
To continue reading
Request your trial