Banzhaf Power Index¶
Power-index model — instantiates Pivotal Participation Leverage Mapping
Measures each participant's voting power as how often their switch is the vote that flips a coalition from losing to winning, counted equally across every possible coalition.
Formal voting weight is not the same thing as real power, and the Banzhaf Power Index is the calculation that exposes the gap. It scores each member by swings: across every possible coalition of the other members, how many times is this member the one whose joining (or leaving) flips the group from losing to winning? A member's Banzhaf power is its share of all such swings. The defining move — and what separates it from its order-based twin, the Shapley–Shubik Power Index — is that coalitions are counted unordered and equally likely: it asks in how many of the possible line-ups am I critical, never in what sequence did we arrive. That makes it the natural instrument for bodies that vote simultaneously yes-or-no, where "who joined first" has no meaning.
Example¶
Three investors hold roughly 45%, 40%, and 15% of a company's voting shares, and a resolution passes on a simple majority. The cap table reads like a clear hierarchy — one big holder, one near-equal, one junior. Banzhaf tells a different story. Enumerate the coalitions: any pair clears 50% (45+40, 45+15, 40+15 all win), but no single holder wins alone. So in each winning pair, both members are critical — remove either and it drops below the line. Count the swings and each of the three is the deciding vote in exactly two coalitions: their Banzhaf power is equal, one-third each. The 15% holder wields precisely as much decisive power as the 45% holder, because winning always takes exactly two and it never matters which two. That single result changes who must be courted — the "junior" investor is a full veto-holder, not a rounding error — and it is invisible to anyone reading only the share percentages.
How it works¶
- Fix the decision rule. Take the members, their weights, and the quota that defines a win.
- Enumerate coalitions. For each of the possible subsets of members, mark it winning or losing under the quota.
- Count swings. A member is critical to a winning coalition if removing it makes the coalition lose. Tally each member's critical memberships across all coalitions.
- Normalize to power. A member's Banzhaf power is its swing count divided by the total swings (relative index), or by the number of other-member coalitions (absolute index).
Because coalitions are weighted equally and order is ignored, the index measures structural decisiveness under the rule — not the influence that ideology or alliances add on top.
Tuning parameters¶
- Quota / decision threshold — the win line the index is scored against; raising it from simple majority to supermajority can make a once-powerful member a dummy, or hand a small member a veto.
- Absolute vs. normalized index — raw swing probability compares power across bodies of different sizes; share-of-total-swings compares members within one body.
- Coalition model — plain yes/no, or a richer space that includes abstention and quorum states; the richer model is truer but multiplies the coalitions to enumerate.
- Player granularity — treat an aligned bloc as one voter or as its separate members; merging or splitting a bloc rewrites every member's index.
- Exact vs. sampled — enumerate all coalitions (exact, feasible for a handful of members) or Monte-Carlo sample them once the count explodes; precision traded for tractability.
When it helps, and when it misleads¶
Its strength is diagnostic: it reveals dummies (members with formal weight but zero swings — no power at all) and disproportionately powerful small members, cutting through the illusion that votes equal influence. Because it is order-free, it fits simultaneous yes/no bodies where the Shapley–Shubik ordering story does not apply.
Its central distortion is the assumption baked into "every coalition equally likely": real members have preferences, allegiances, and enmities, so many counted coalitions never actually form. Taken literally over a polarized body, the index can badly misstate practical power. It also grows combinatorially, so large bodies force sampling. The classic misuse is computing it once and treating the number as destiny — or running it backwards to rationalize a weighting already chosen. The discipline is to read it as a structural baseline under a stated assumption, overlay preference and alliance information before acting, and recompute whenever the membership or the quota moves.[1]
How it implements the components¶
power_index_metric— it is the metric: a per-member power score derived from swing frequency.marginal_contribution_estimate— each swing is a marginal contribution (the member is the margin that turns a loss into a win); the index aggregates these into a single figure.
It does not define the decision rule and roster it runs on — those come from Minimal Winning Coalition Enumeration and the body's charter — nor test a single named actor's live removal (that is Dependency Removal Counterfactual); and its order-weighted counterpart, the Shapley–Shubik Power Index, scores power over arrival orderings rather than unordered coalitions.
Related¶
- Instantiates: Pivotal Participation Leverage Mapping — Banzhaf supplies a quantified power score for the map.
- Consumes: the decision rule and weights — supplied by the body's charter or by Weighted Voting Simulation.
- Sibling mechanisms: Minimal Winning Coalition Enumeration · Shapley–Shubik Power Index · Weighted Voting Simulation · Swing-Vote Scenario Review · Pivotality Counterfactual Matrix · Dependency Removal Counterfactual · Veto-Point Review
Editorial Notes¶
Form Classification¶
Form family: Analysis, Modeling & Optimization
Rationale: Measures each participant's voting power as how often their switch is the vote that flips a coalition from losing to winning, counted equally across every possible coalition, making its operative form a computation or analytic transformation that produces an inference, comparison, or optimized result.
Independent corroboration: The frozen evidence defines Banzhaf Power Index as 'Measures each participant's voting power as how often their switch is the vote that flips a coalition from losing to winning, counted equally across every possible coalition', so its operative form is Analysis, Modeling & Optimization.
Review outcome: Independent reviewer agreement; high confidence.
Origin Attribution¶
Primary origin: Mathematics
Origin pattern: Single lineage
Present-day reach: Specialized
Rationale: Banzhaf's combinatorial swing-counting formalized voting power as pivotality across equally counted coalitions.
Related originating lineages:
- Economics & Finance — Cooperative game theory situates the index among formal power and allocation measures.
- Political Science — Weighted legislatures and collective decision bodies provide the core application.
Review outcome: Independent reviewer agreement; high confidence.
References¶
[1] The index comes from John F. Banzhaf III's 1965 analysis of weighted voting, which showed that on real bodies such as the Nassau County Board of Supervisors some members held large weighted votes yet were never mathematically pivotal — dummies with zero power. That formal weight and real power diverge is the whole reason to compute it. withdrawn registry ↩