Source-linked AI summary
Who Pays for a Connected Public Good?
Marco Tulio Angulo
TL;DR
The paper asks how network connections constrain voluntary payment for a public good that benefits participants and nonparticipants. It models participation and committed cost sharing using output from the largest connected active group. It proves that the path maximizes financeable cost among connected networks with at least five players, while minimizing surviving output after a random withdrawal, and that maximal capacity approaches one quarter of full-participation gross benefit.
Problem
When only participants pay for a public good that also benefits nonparticipants, it is unclear how network position constrains feasible payments and the cost a group can finance.
Source: §1 · pp. 1–2Method
The paper studies a participation game where output is the size of the largest connected active group and committed charges divide a fixed operating cost.
Source: §1 · p. 2, §3.1 · p. 8Results
Among connected networks with at least five players, the path uniquely maximizes cost-bearing capacity and minimizes expected surviving output after one equally likely withdrawal; its maximal capacity approaches one quarter of full participation’s gross social benefit.
Source: §1.1 · p. 4Takeaways & Limitations
Fragility can strengthen voluntary finance, but even the most financeable network cannot close the gap between the value created and what participants will pay to sustain it.
Source: §4 · p. 19Supportability is an existence result under a committed, budget-balanced schedule and does not predict which schedule, group, or equilibrium will emerge.
Source: §4 · p. 18
Abstract
from arXiv · showhide
Connectivity can turn separate contributions into a public good that benefits participants and nonparticipants alike. Yet only participants pay for it, and each can avoid her charge by withdrawing while enjoying whatever public output survives. How do network connections constrain who can or must pay, and how large a cost can the group bear? We study a participation game in which public output equals the size of the largest connected group of active players, whose members divide a fixed operating cost. The output a participant destroys by withdrawing is her withdrawal responsibility and bounds her payment. These bounds determine the group's cost-bearing capacity and, whenever the group can be kept active, which participants must pay. We prove that among connected networks with at least five players, the path uniquely maximizes the cost-bearing capacity of a group, yet minimizes the expected surviving output after a uniformly selected participant withdraws. As the group grows, the ratio of maximal capacity to gross social benefit under full participation converges to one quarter. Fragility can therefore strengthen voluntary finance, but even the most financeable network cannot close the gap between the value it creates and what its participants will pay to sustain it.
1. Introduction
The introduction frames voluntary finance as a connected-public-good problem: participants pay while everyone benefits, and a participant’s withdrawal can reduce or fragment output. It presents a committed cost-sharing game and relates the results to network vulnerability, public-good provision, and cost allocation.
1. Introduction
- The motivating settings include emergency radio relays, community broadband, and connected habitat conservation, where contributors pay but benefits extend beyond them.A participant’s network position affects whether withdrawal removes one contribution or fragments a larger system.Source: §1 · pp. 1–2
1.1. Summary of contributions
- Each participant’s withdrawal loss bounds her charge; the resulting payment floors and ceilings determine feasible divisions and the group’s cost-bearing capacity.A participant can be exempted when others can cover the bill; at maximum capacity, the payment division is unique.Source: §1.1 · p. 3
- Among connected networks with at least five participants, the path uniquely maximizes cost-bearing capacity and minimizes expected surviving output after one equally likely withdrawal.For eight participants, the path can bear a bill of twenty times the unit benefit versus eight for the cycle, despite equal initial output.Source: §1.1 · pp. 3–4
- For groups of at least three, every participant can be exempted at bills no greater than half of capacity, but higher bills force the largest withdrawal-loss participants to pay first.A hub-and-spoke network forces its hub to pay just above half of capacity, while a growing path delays its first forced payer until near capacity.Source: §1.1 · p. 4
- Even when full participation maximizes welfare, no feasible bill division may sustain it; the path’s maximal financeable bill approaches one quarter of full participation’s gross social benefit.A withdrawing participant counts her own lost benefit, while her departure reduces benefits for all players.Source: §1.1 · p. 4
1.2. Related work
- The model makes links productive rather than reciprocal or consumer-benefit links, and makes participation endogenous while holding infrastructure fixed and dividing one operating bill.Unlike external deletion models, voluntary withdrawal responds to a charge; the paper aggregates withdrawal responsibility across vertices, while noting the vertex statistic itself is not new.Source: §1.2 · pp. 5–6
2. The model
The model represents players as positions on a fixed connected graph and lets each choose whether to participate. All players receive benefit from the largest connected component of active players, while active players divide a fixed operating cost under a committed payment rule.
2. The model
- Players simultaneously choose active or inactive status on a finite, simple, undirected, connected graph; any vertex subset can be an active set.The model permits disconnected active sets, and a withdrawal can fragment a connected set.Source: §2 · p. 6
- Public output is the size of the largest connected component induced by the active set, and both participants and outsiders receive its public benefit.The unit benefit is b; the empty active set produces zero output.Source: §2 · pp. 6–7
- Every nonempty active set incurs the same real operating cost, which participants must cover through nonnegative, budget-balanced charges.The institution commits to the payment rule before activity choices, and a unilateral entry or withdrawal invokes the charges specified for the resulting active set.Source: §2 · p. 7
- An active set is a Nash equilibrium when no player benefits by changing activity, and it is strict when every player would lose by doing so.Active players compare benefit net of their charge; outsiders receive the public benefit without paying.Source: §2 · p. 7
3. Results
The results link network position to each participant’s payment limit, the group’s financing capacity, and which participants must pay. They show that the path maximizes capacity but can leave a gap between socially valuable full participation and what participants can finance.
3.1. Participation incentives and Nash supportability
- A connected group’s payment ceiling is the value of output a participant’s withdrawal destroys; a proper group must also deter entry by an adjacent outsider.Full participation is supportable when C≤bR(V); a proper connected group requires b≤C≤bR(A).Source: §3.1 · pp. 8–10
3.2. Who can pay, who can be exempted, and who must pay?
- A participant can be exempted iff C≤b[R(A)−ri(A)]; she must pay otherwise, while bearing the whole bill is possible iff C≤bri(A).When others cannot cover the bill, her minimum payment is their uncovered shortfall.Source: §3.2 · p. 11
3.3. Network architecture and cost-bearing capacity
- For m≥5, the path uniquely maximizes cost-bearing capacity and minimizes expected surviving connected output after one uniformly selected withdrawal.The path’s high capacity lets the institution exempt any chosen participant over a wide range of supportable costs.Source: §3.3 · p. 14
3.4. Paying no more for greater responsibility
- Under the restriction that higher responsibility cannot mean a higher charge, capacity falls from bF_m to bm on paths with m≥3.The restriction reduces capacity whenever R(A)>m.Source: §3.4 · p. 15
3.5. Social welfare and the voluntary-financing gap
- When bR(V)<C<bN^2, full participation uniquely maximizes welfare but no rule sustains it; even the path finances asymptotically only one quarter of gross benefit.The gap reflects that a participant counts only her own lost benefit, while her withdrawal reduces every player’s benefit.Source: §3.5 · p. 16
4. Discussion and conclusion
Network position sets the payment limits created by voluntary participation, while institutions choose among feasible cost divisions. The discussion connects this tradeoff to welfare, benchmark assumptions, and directions for extending the analysis.
Payment constraints
- A participant’s payment lies between a network-determined floor and ceiling; only at cost-bearing capacity do all payment ranges collapse to one division.The ceiling is the connected output her withdrawal destroys; the floor is zero if the others can cover the bill, otherwise their shortfall.Source: §4 · p. 17
Architecture and finance
- For a fixed group size, output, benefits, and cost, resilient links can reduce charges participants will bear, while concentrated dependence can raise capacity and reduce expected surviving output.The comparison concerns one uniformly selected withdrawal, not targeted, repeated, or correlated failures, and is not a prescription to build fragile networks.Source: §4 · p. 17
Payment rules
- Extra capacity from larger withdrawal losses requires permitting higher charges for participants with greater losses; nonincreasing-charge rules forgo that capacity.Withdrawal loss is a participation constraint, not a moral responsibility, bargaining power, or entitlement; institutions and social norms select among feasible divisions.Source: §4 · p. 18
Welfare implications
- Voluntary payments may fail to sustain full participation even when total benefits exceed operating cost, and the path’s financeable share approaches one quarter of gross social benefit.The path maximizes capacity, but still cannot close the gap between created value and what participants will pay to sustain it.Source: §4 · p. 18
Scope of supportability
- Supportability is an existence result under a committed, nonnegative, budget-balanced, identity-dependent charge schedule; it does not predict the chosen schedule, group, or equilibrium.Off-equilibrium charges may not be credible after renegotiation, inactivity may coexist with provision, and participants are only weakly willing to remain at maximum capacity.Source: §4 · p. 18
Extensions and scope
- Extensions include variable costs, weighted or directed links, endogenous networks, sequential participation, empirical tests of withdrawal losses, and bargaining over feasible divisions.The benchmark conclusions are specific to fixed, known, unweighted, undirected networks, binary simultaneous participation, common linear benefits, uniform nonempty-set costs, and participant-only payments.Source: §4 · pp. 18–19
Appendix A. Proof of the sharp aggregate-responsibility bound
The appendix proves a sharp upper bound on aggregate withdrawal responsibility and classifies the connected graphs attaining it. For groups of at least five, only a path attains the bound.
Proof strategy
- Spanning-tree domination reduces the upper-bound proof to trees, where a centroid splits the tree into branches of at most half its order and rooted-subtree bounds constrain their total.The tree bound’s equality conditions force the extremal paths, with the four-vertex star as the additional small equality case.Source: Appendix A · pp. 20–22
Lower bound
- Deleting any vertex from a connected graph leaves at most m−1 vertices, so every responsibility is at least one; equality throughout holds exactly when every deletion leaves the graph connected.Source: Appendix A · p. 21
Path construction
- On Pm, vertex i has responsibility min{i, m + 1 − i}, giving the endpoint-rooted path’s triangular sequence of responsibilities.The rooted-subtree bound has equality precisely for a path rooted at an endpoint.Source: Appendix A · pp. 20, 23
Sharp bound and equality cases
- For connected graphs of order m≥5, the path Pm uniquely attains the maximum aggregate withdrawal responsibility; P3, and P4 and K1,3, are the smaller equality cases.The proof rules out extra edges by showing they would lower at least one vertex’s withdrawal responsibility.Source: Appendix A · pp. 23–24
Appendix B.1. Disconnected active sets
The appendices extend payment-support results to disconnected groups and group-level payments, then characterize capacity under ordered charges and introduce local payment certificates when participants have partial network information.
Appendix B.1. Disconnected active sets
- For a proper active set, strict support requires positive responsibility for every participant and bqmax(A)<C<bR(A); at full participation it requires positive responsibilities and 0≤C<bR(V).Source: Appendix B.1 · p. 24
Appendix B.2. Payments by groups
- A subgroup T can bear the entire operating cost exactly when C≤brA(T), a special case of the stated feasible interval for its total payment.Source: Appendix B.2 · p. 25
Appendix B.3. Responsibility-nonincreasing payments
- For a connected group of m≥2, responsibility-nonincreasing charges support full participation for 0≤C≤bm and a proper set for b≤C≤bm; strict support tightens the relevant bounds.Source: Appendix B.3 · p. 26
Appendix C. Partial information and local responsibility certificates
- With only nearby network information, the appendix derives the largest payment ceiling justified by a bounded view; it does not study communication or distributed computation.Remote links matter because they determine whether branches reconnect after a participant withdraws.Source: Appendix C · pp. 26–27
Appendix C.1. Local responsibility certificates
The local responsibility certificate is the strongest lower bound justified by a participant’s bounded rooted neighborhood and known group size. It matches exact responsibility when the observation contains the whole active network.
Information boundary
- The observation does not reveal links leaving its frontier, so apparent branch separation may not establish that branches remain disconnected after withdrawal.Components not meeting the frontier are treated as sealed in the certificate construction.Source: Appendix C.1 · p. 27
Certificate guarantee
- The certificate is the largest lower bound on withdrawal responsibility valid for every connected m-vertex network consistent with a participant’s rooted k-step observation.Source: Appendix C.1 · p. 27
Exactness
- The certificate equals exact withdrawal responsibility when the observed neighborhood contains the entire active network.Source: Appendix C.1 · p. 27
Radius
- The certificate is monotone with observation radius: the bound from a larger neighborhood cannot be lower than the one from a smaller neighborhood.Source: Appendix C.1 · p. 28
Appendix C.2. Cost-bearing capacity from local views
Local cost-bearing capacity limits the bill supportable from individual certificates, without requiring the institution to pool network maps. For paths, a small observation radius relative to group size certifies a vanishing share of full-information capacity.
Appendix C.2. Cost-bearing capacity from local views
- The institution can cover a bill up to the sum of participants’ certified payment ceilings, but must collect the certificates to divide it.It need not pool the underlying network maps; pooling them could certify a larger bill.Source: Appendix C.2 · pp. 28–29
- At full participation, local-view capacity is the largest Nash-supportable cost under the certification requirement; a proper connected group also needs cost at least b and the proof’s post-entry charges.Source: Appendix C.2 · p. 29
- On a path with m ≥ 2k + 2, vertices k−1 steps from either endpoint have certificates 2 through k, while every other certificate equals one.Source: Appendix C.2 · p. 29
- When k/m tends to zero, C(k)(A)/[bR(A)] tends to zero, so certifying a nonvanishing share of full-information path capacity requires an observation radius proportional to group size.The same argument applies to verified lower bounds on withdrawal responsibility.Source: Appendix C.2 · p. 29
Appendix D.1. Coordinated withdrawal
Individual payment ceilings prevent a coalition of participants from benefiting by withdrawing together, even when members can transfer money among themselves. The guarantee holds with other actions and the cost-division rule fixed, and excludes deviations mixing entry and withdrawal.
Appendix D.1. Coordinated withdrawal
- At a Nash equilibrium, no nonempty coalition of participants can improve its combined payoff by jointly withdrawing, even with budget-balanced transfers.If the equilibrium is strict, joint withdrawal strictly reduces the coalition’s combined payoff.Source: Appendix D.1 · pp. 29–30
- Each departing member loses at least the public benefit tied to her individual withdrawal responsibility, so the coalition’s combined gain is nonpositive.Source: Appendix D.1 · p. 30
- The result holds with other players’ actions and the cost rule fixed; it does not cover strong-Nash deviations or deviations mixing entry and withdrawal.Source: Appendix D.1 · p. 30
Appendix D.2. Nonlinear and heterogeneous public benefits
With nondecreasing, heterogeneous public benefits, participants’ withdrawal values bound their charges while prospective entrants’ incentives set a lower bound on charges. The appendix also treats immutable personal costs, which cannot be reallocated among participants.
Appendix D.2. Nonlinear and heterogeneous public benefits
- With nondecreasing heterogeneous benefits, Nash support requires participants’ charges not to exceed their withdrawal values and prospective entrants’ charges to meet their entry incentives.Current participants set the upper bound, while prospective entrants set the lower bound.Source: Appendix D.2 · p. 30
- In the generalized model, participants’ charge ceilings also prevent coordinated withdrawal from increasing a coalition’s combined payoff.If every participant’s withdrawal value is positive, supporting rules have a unique charge division minimizing the largest charge-to-withdrawal-value ratio.Source: Appendix D.2 · p. 31
Appendix D.3. Immutable personal costs
- With immutable personal costs, an active set is a Nash equilibrium exactly when each active player’s cost κ_i is at most b r_i(A) and each outsider’s is at least b q_j(A).Source: Appendix D.3 · p. 31
- The immutable-cost benchmark is an exact potential game, unlike the main model’s reallocated fixed operating cost; the main capacity result depends on the institution’s ability to divide C.Source: Appendix D.3 · p. 31