Robust rational turnout (Q836880): Difference between revisions
From MaRDI portal
Removed claim: reviewed by (P1447): Item:Q283988 |
Changed an Item |
||
Property / reviewed by | |||
Property / reviewed by: Tadeusz Radzik / rank | |||
Normal rank |
Revision as of 13:34, 12 February 2024
scientific article
Language | Label | Description | Also known as |
---|---|---|---|
English | Robust rational turnout |
scientific article |
Statements
Robust rational turnout (English)
0 references
9 September 2009
0 references
The author considers a two-candidate voting game in which the set \(N\) of voters (players) is an electorate split into two disjoint groups \(N = N_1\cup N_2\), where \(N_1=\{1,\dots, n_1\}\) and \(N_2=\{n_1+1,\dots, n_1+n_2\}\) for some naturals \(n_1\) and \(n_2\) with \(n_1+n_2\geq 2\). For \(t=1,2\), it is assumed that each member \(i\in N_t\) can only support candidate number \(t\) at a constant cost \(c>0\) or abstain. The strategy of voter \(i\) is represented by \(\sigma_i\in [0, 1]\) where \(\sigma_i\) is the probability that this player votes, and let \(\sigma\in [0, 1]^{n_1+n_2}\) represent a vector consisting of all players' voting strategies. Let \(m_1\) and \(m_2\) be the actual total votes cast by groups \(N_1\) and \(N_2\), respectively. Candidate 1 wins the election if \(m_1+m_2-k>0\) and candidate 2 wins the election if \(m_1+m_2-k<0\), where \(k\) is some fixed integer. When \(m_1+m_2-k=0\), candidate 1 wins with probability \(\pi\in \{1, \frac{1}{2}\}\) (two cases for \(\pi\) are considered), and candidate 2 wins with probability \(1-\pi\). For \(t=1,2,\) player \(i\in N_t\) receives benefit 1 when candidate \(t\) wins the election, and 0, otherwise. This game with common cost of voting is described as \(\Gamma(c,\pi, k,n_1, n_2)\). Three types of equilibria \(\sigma\) in the game \(\Gamma\) are considered: Nash equilibria, quasistrict Nash equilibria and regular Nash equilibria. The first result by the author shows that in every Nash equilibrium the behavior of indifferent players is characterized by at most two distinct probabilities. The second result says that almost all games \(\Gamma\) have their Nash equilibria regular, and the third one gives some relationship between game \(\Gamma\) and a set set of its perturbations. The last result obtained in the paper concerns the robustness of the equilibrium set to the introduction of two forms of incomplete information. One is with regard to individual cost of voting, and the second concerns the size of the two groups of voters where the voters are also admitted to vote for their non-favorite candidates.
0 references
participation game
0 references
voting game
0 references
regular Nash equilibrium
0 references
incomplete information
0 references