The inevitability of the "paradox of redistribution" in the allocation of voting weights

Dietrich Fischer, Andrew Schotter

    Research output: Contribution to journalArticle

    Abstract

    In voting bodies, when voting weights are reallocated, it may be observed that the voting power of some members, as measured by the Shapley-Shubik and Banzhaf power indices, increases while their voting weight decreases. By a simple constructive proof, this paper shows that such a "paradox of redistribution" can always occur in any voting body if the number of voters, n, is sufficiently large. Simulation results show that the paradox is quite frequent (up to 30 percent) and increases with n (at least for small n). Examples are given where the Banzhaf and Shapley-Shubik indices are not consistent in demonstrating the paradox.

    Original languageEnglish (US)
    Pages (from-to)49-67
    Number of pages19
    JournalPublic Choice
    Volume33
    Issue number2
    DOIs
    StatePublished - Sep 1978

    Fingerprint

    redistribution
    voting
    Redistribution
    Voting
    Paradox
    simulation

    ASJC Scopus subject areas

    • Sociology and Political Science
    • Economics and Econometrics

    Cite this

    The inevitability of the "paradox of redistribution" in the allocation of voting weights. / Fischer, Dietrich; Schotter, Andrew.

    In: Public Choice, Vol. 33, No. 2, 09.1978, p. 49-67.

    Research output: Contribution to journalArticle

    Fischer, Dietrich ; Schotter, Andrew. / The inevitability of the "paradox of redistribution" in the allocation of voting weights. In: Public Choice. 1978 ; Vol. 33, No. 2. pp. 49-67.
    @article{ccd38a15356c484ab31b4c14aafeb60b,
    title = "The inevitability of the {"}paradox of redistribution{"} in the allocation of voting weights",
    abstract = "In voting bodies, when voting weights are reallocated, it may be observed that the voting power of some members, as measured by the Shapley-Shubik and Banzhaf power indices, increases while their voting weight decreases. By a simple constructive proof, this paper shows that such a {"}paradox of redistribution{"} can always occur in any voting body if the number of voters, n, is sufficiently large. Simulation results show that the paradox is quite frequent (up to 30 percent) and increases with n (at least for small n). Examples are given where the Banzhaf and Shapley-Shubik indices are not consistent in demonstrating the paradox.",
    author = "Dietrich Fischer and Andrew Schotter",
    year = "1978",
    month = "9",
    doi = "10.1007/BF00118357",
    language = "English (US)",
    volume = "33",
    pages = "49--67",
    journal = "Public Choice",
    issn = "0048-5829",
    publisher = "Springer Netherlands",
    number = "2",

    }

    TY - JOUR

    T1 - The inevitability of the "paradox of redistribution" in the allocation of voting weights

    AU - Fischer, Dietrich

    AU - Schotter, Andrew

    PY - 1978/9

    Y1 - 1978/9

    N2 - In voting bodies, when voting weights are reallocated, it may be observed that the voting power of some members, as measured by the Shapley-Shubik and Banzhaf power indices, increases while their voting weight decreases. By a simple constructive proof, this paper shows that such a "paradox of redistribution" can always occur in any voting body if the number of voters, n, is sufficiently large. Simulation results show that the paradox is quite frequent (up to 30 percent) and increases with n (at least for small n). Examples are given where the Banzhaf and Shapley-Shubik indices are not consistent in demonstrating the paradox.

    AB - In voting bodies, when voting weights are reallocated, it may be observed that the voting power of some members, as measured by the Shapley-Shubik and Banzhaf power indices, increases while their voting weight decreases. By a simple constructive proof, this paper shows that such a "paradox of redistribution" can always occur in any voting body if the number of voters, n, is sufficiently large. Simulation results show that the paradox is quite frequent (up to 30 percent) and increases with n (at least for small n). Examples are given where the Banzhaf and Shapley-Shubik indices are not consistent in demonstrating the paradox.

    UR - http://www.scopus.com/inward/record.url?scp=34248013756&partnerID=8YFLogxK

    UR - http://www.scopus.com/inward/citedby.url?scp=34248013756&partnerID=8YFLogxK

    U2 - 10.1007/BF00118357

    DO - 10.1007/BF00118357

    M3 - Article

    AN - SCOPUS:34248013756

    VL - 33

    SP - 49

    EP - 67

    JO - Public Choice

    JF - Public Choice

    SN - 0048-5829

    IS - 2

    ER -