Beruflich Dokumente
Kultur Dokumente
!
Public Choice and Arrow’s Impossibility
Theorem: Implications for the Public Policy
Discipline
Gilbert Michaud
!
22 Virginia Policy Review
I. Public Choice
!
Similarly, electorates back propositions and candidates that
they believe will maximize their respective utility (McLean,
1991). However, voters in democracies face a small probability
that their vote will decidedly change the result of the elections,
and collecting the information necessary for a well-informed
voting decision requires substantial time and work. Thus,
public choice theory claims it to be rational for voters to be
largely uninformed of government and politics and perhaps
even refrain from voting (Downs, 1998). Downs’ ‘paradox of
voting’ observation may explain low voter turnout.
!
24 Virginia Policy Review
!
Most economists today agree that individuals have ordinal
preferences, rather than a numeric scale of psychic happiness
that one can gather from goods, services, actions, etc. (i.e.,
cardinal preferences) (Samuelson & Nordhaus, 2009). In other
words, a combination of goods (or candidates) can be ordered
such that each is measured as better, worse, or equal to
another. However, it is uncertain how much more or less these
individuals might prefer a good or candidate in comparison
with another (Samuelson & Nordhaus, 2009). Each individual’s
ordinal favorite forms a preference relation, which provides
information to determine collective decisions. How can the
honest revelation of these preference orderings be assured? If
individuals cannot agree with how to rank a candidate, then
how can it be claimed that society as a whole ranks these
outcomes one way or the other?
!
region, which is populated by 1,000 voters. Now, assume that
(R) has 450 voters who would elect him, but would instead
vote for (D) if (R) were not running. (D), on the other hand,
has 350 voters who back him primarily, but who rank (R) as
their second choice. Lastly, (L) has only 200 voters who are
committed supporters, and they would vote for (D) if (L)
withdraws, but not for (R). Hence, the voters’ ranking
preference may be presented as follows: 450 voters
(R)>(D)>(L); 350 voters (D)>(R)>(L); 200 voters
(L)>(D)>(R). What is seen is that of the non-(R) voters, a clear
majority of 350 out of 550 backs the (D) candidate. An
overwhelming majority of non-(R) voters would vote for (R) if
(D) decides to withdraw. However, presenting all three
candidates makes the election less democratic, particularly due
to the lack of political equality in the voting process.
If all three candidates run, the (R) simply wins with 450 votes
(scenario 1). However, suppose the (L) candidate withdraws.
All of its 200 supporters vote for their next choice, the (D)
candidate. As a result, (D) wins with 550 votes to 450 votes
(scenario 2). In scenario 1, (R) won despite the fact that they
were one of the top two choices of only 800 voters. The (D)
lost even though they were one of the top two choices of all
1,000 voters. Hence, the resulting outcome is paradoxical since
adding more ballot choices actually results in a less equitable
outcome. Here, a candidate who ranked first or second on
everyone’s list loses to a candidate who has lesser support in
the electorate’s first or second place ranking. Why is it that the
(R) candidate wins in one scenario and the (D) candidate wins
in the other, although the preferences of the citizenry remained
the same?
!
26 Virginia Policy Review
!
Once again, there is a sub-optimal result where the winning
candidate is the choice of only 800 voters, while a non-running
candidate (D) ranks in the top two in everyone’s preference
list. What scenario 3 demonstrates is that if either opposing
candidate in an election has to withdraw, the less popular one
should consider giving way to the favorite in order to ensure an
outcome more reflective of the constituency. Of course, this is
hardly applicable in the real world. Contrary to common belief,
restricting voters’ choices by having one opposition candidate
withdraw could produce a more democratic conclusion.
!
referenda) may be inaccurate. They must also note the fallacies
in the plurality voting method, and perhaps explore ordinal
ranking schemes, such as the Borda count system, as an
alternative approach to determining a winner.
III. Conclusion
!
28 Virginia Policy Review
!
as the lead researcher for the Energy & Power segment, producing over forty
organizational case studies.
Virginia Policy Review 29
!
References
!
30 Virginia Policy Review
!
McLean, I. (1991). Public choice: An introduction. New York, NY:
Basil Blackwell.