Enter the total number of players (n) and either the number of permutations in which a given player is pivotal (Pi) or the Shapley Shubik Power Index (SSPI) into the calculator to determine the missing value.
Related Calculators
- Banzhaf Index Calculator
- Permutation Multiplication Calculator
- Binomial Coefficient Calculator
- Effective Number of Parties Calculator
- Gallagher Index Calculator
- All Math and Numbers Calculators
Shapley Shubik Power Index Formula
The following formula is used to calculate the Shapley Shubik Power Index for player i in a voting game.
SSPI_i = P_i / n!
Variables:
- SSPIi is the Shapley Shubik Power Index of player i.
- Pi is the number of permutations (orderings) of all players in which player i is pivotal (i.e., the first player in the order whose entry turns a losing set of earlier players into a winning coalition).
- n is the total number of players.
- ! denotes factorial (e.g., n! = 1x2x3x…xn, with 0! = 1).
What is the Shapley Shubik Power Index?
The Shapley Shubik Power Index (SSPI) is a method from cooperative game theory for measuring how much actual voting influence a player holds in a weighted voting system. The core insight is that vote share and voting power are not the same thing. A player with 30% of votes may hold 60% of the real decision-making power, or conversely, may be nearly irrelevant, depending on how their votes combine with others to form winning coalitions.
The SSPI captures this by imagining all possible orderings in which players might cast their support. In each ordering, one player is identified as pivotal: the voter whose inclusion pushes the running total over the required threshold (the quota). The SSPI for a player is simply the fraction of all possible orderings in which that player is the pivotal vote. Because exactly one player is pivotal in each ordering, the SSPI values across all players always sum to exactly 1.
This ordering-based perspective is what distinguishes the SSPI from simpler vote-count measures. A player who holds a veto, or whose votes are frequently the deciding factor in crossing the quota threshold, will score much higher than their raw vote percentage would suggest. A player who is either always on the winning side before they are needed, or always insufficient to push any coalition over the threshold, approaches a power index of zero regardless of how many nominal votes they hold.
Historical Background and Academic Origins
The Shapley Shubik Power Index was introduced in a landmark 1954 paper by Lloyd S. Shapley and Martin Shubik, titled “A Method for Evaluating the Distribution of Power in a Committee System,” published in the American Political Science Review. The paper adapted Shapley’s earlier (1953) cooperative game theory value to the specific structure of simple yes-or-no voting games, creating the first mathematically rigorous framework for quantifying political power.
Lloyd Shapley received the Nobel Memorial Prize in Economic Sciences in 2012 (shared with Alvin Roth) in part for this foundational work in game theory. Martin Shubik spent most of his career at Yale, where he continued to develop the application of game theory to political economy. Their 1954 collaboration remains one of the most cited papers in the intersection of mathematics and political science, with thousands of academic references spanning seven decades.
The motivation behind the original paper was a practical observation: legislative bodies routinely assign different vote weights to members, but there was no principled way to translate those formal weights into a measure of actual influence. The SSPI solved this by grounding power measurement in the logic of pivotal players across all possible coalition-formation sequences.
Power Distribution in Real Voting Bodies: Documented Data
The SSPI produces dramatically non-intuitive results when applied to real-world bodies. The following data illustrates the gap between nominal vote share and SSPI-measured power.
UN Security Council: A 105-to-1 Power Gap
The UN Security Council has 15 members: 5 permanent members (USA, Russia, China, France, UK) and 10 rotating non-permanent members. A resolution requires at least 9 affirmative votes including all 5 permanent members, meaning each permanent member holds an absolute veto. The SSPI analysis of this structure produces a stark result across the approximately 1.3 trillion possible orderings of 15 members.
| Member Type | Count | Nominal Vote Share | SSPI per Member | Total Group SSPI |
|---|---|---|---|---|
| Permanent Member (P5) | 5 | 6.67% | 19.63% (421/2145) | 98.1% |
| Non-Permanent Member | 10 | 6.67% | 0.19% (4/2145) | 1.9% |
The power ratio between a permanent and non-permanent member is approximately 105:1, despite both holding a single nominal vote. The ten non-permanent members collectively hold less than 2% of the total SSPI-measured power. This asymmetry arises directly from the veto structure: a non-permanent member can only be pivotal in the rare ordering where exactly the right subset of permanent members has already voted yes, leaving a gap of exactly one vote that the non-permanent member fills. In most orderings, this condition does not occur.
Weighted Voting Example: [6: 4, 3, 2]
Consider a 3-player weighted voting game with quota 6, where Player A holds 4 votes, Player B holds 3, and Player C holds 2 (total = 9 votes). There are 3! = 6 possible orderings. In 4 orderings, Player A is pivotal; in 1 ordering, Player B is pivotal; in 1 ordering, Player C is pivotal.
| Player | Votes | Vote Share | Pivotal Orderings (Pi) | SSPI | Power vs. Vote Share |
|---|---|---|---|---|---|
| Player A | 4 | 44.4% | 4 of 6 | 66.7% | +22.3 points |
| Player B | 3 | 33.3% | 1 of 6 | 16.7% | -16.6 points |
| Player C | 2 | 22.2% | 1 of 6 | 16.7% | -5.6 points |
Player A receives 44.4% of votes but commands 66.7% of the SSPI-measured power. Players B and C, despite having different vote counts, share identical power indices because they are equally often pivotal. Two players with different vote totals can have exactly the same real influence in a game, which is a result that no simple vote-counting method can detect.
SSPI vs. Banzhaf Power Index: When They Agree and When They Differ
The Shapley Shubik Power Index and the Banzhaf Power Index are the two dominant frameworks for measuring voting power, and they frequently produce different answers from identical voting structures. The Banzhaf index counts the number of coalitions in which a player is a critical member (removal flips the coalition from winning to losing), treating all such coalitions as equally likely. The SSPI instead counts orderings (permutations) in which a player is pivotal, treating all orderings as equally likely. The conceptual difference is whether one models voting as simultaneous coalition formations (Banzhaf) or as a sequential process where players arrive one at a time (Shapley Shubik).
In practice, the two indices produce identical results for symmetric games where all players have equal weights. They diverge most sharply in games with strong asymmetries, such as bodies with veto players or very unequal vote distributions. The SSPI is generally more appropriate when coalition formation is genuinely sequential or ordered; the Banzhaf index better reflects situations where coalitions form independently and simultaneously. Both have been applied to EU Council voting analysis across treaty eras, often producing meaningfully different power rankings for mid-sized member states.
The Dummy Voter Phenomenon
One of the most counterintuitive results produced by the SSPI is the existence of dummy voters: players who hold a positive number of formal votes but whose SSPI is exactly zero. Consider the voting rule [5: 3, 3, 1]. The quota is 5, and there are three players holding 3, 3, and 1 vote respectively. The only winning coalition involving Player C (with 1 vote) would be {A, B, C} with 7 votes, but {A, B} already wins with 6 votes without needing C. There is no losing coalition that Player C alone could convert into a winning one. Player C is therefore a dummy voter with SSPI = 0, despite holding a valid vote.
Dummy voters arise in real-world bodies when thresholds are set such that one group can always reach the quota without involving smaller members. The SSPI provides the formal diagnostic tool to identify these situations, which is one reason it was adopted in the analysis and design of international organizations. A reformed voting rule that eliminates dummy voters distributes power more equitably, even if the nominal vote allocations change only marginally.
Mathematical Properties and Axioms
The Shapley Shubik Power Index satisfies four axiomatic properties that together uniquely characterize it as a power measure. These were used by Shapley and Shubik to derive the index as the unique measure satisfying all of them simultaneously.
- Efficiency: The sum of all players’ SSPI values equals exactly 1. No power is created or destroyed; the index partitions total influence among all players.
- Symmetry: If two players are interchangeable in the game (swapping them leaves all winning coalitions unchanged), they receive identical SSPI values. Power is determined solely by strategic position, not identity.
- Dummy: A player who is never pivotal in any coalition receives an SSPI of zero. Nominal vote ownership does not guarantee influence.
- Additivity: If a game can be decomposed into independent sub-games, a player’s power in the combined game equals the sum of their power in each sub-game. This enables the index to scale consistently across complex multi-stage voting structures.
Computational Complexity
Computing the SSPI exactly requires enumerating all n! permutations, which grows factorially: 10 players require 3.6 million orderings, 15 players require 1.3 trillion, and 20 players require approximately 2.4 quintillion. Direct enumeration becomes computationally infeasible for large bodies. For weighted majority games specifically, dynamic programming algorithms reduce computation to pseudo-polynomial time O(n x W), where W is the total weight sum, making exact computation tractable for bodies like the EU Council. For general cooperative games without the weighted majority structure, computing exact Shapley values is #P-complete, meaning no polynomial-time algorithm is known. In practice, Monte Carlo sampling is used for large games: a sample of 10 million random permutations for n = 100 players typically yields estimates accurate to within 0.5 percentage points with high probability.
Applications Across Domains
In corporate governance, the SSPI measures the real influence of shareholders in companies with complex ownership structures, dual-class shares, or supermajority requirements. A shareholder with a 15% stake can have dramatically more or less power than their stake implies, depending on the distribution of remaining shares and the decision threshold.
In international treaty organizations, the index has been applied to the EU Council of Ministers through multiple treaty eras (Rome, Maastricht, Nice, Lisbon), tracking how enlargement rounds and treaty reforms shifted power among member states. Germany, France, and other large members have seen SSPI values fluctuate substantially as quota rules and member counts changed.
In electoral college systems, the SSPI formalizes the mechanism behind the well-documented strategic over-weighting of large swing states. Because a unit-rule state awards all its electoral votes to one candidate, larger states can have disproportionate SSPI values relative to their population share.
Emerging research has extended SSPI-style analysis to multi-level governance (where regional bodies vote within national structures which vote within international structures), blockchain governance (measuring token-holder influence in on-chain voting protocols), and AI governance (analyzing weighted decision-making in multi-stakeholder AI safety boards where different parties hold different vote allocations).
