PASS Theorem 1: own-account pay exists iff no cheaper plan replicates every account record law (500 LPs, 196 infeasible, with and without limited liability)
PASS Theorem 1 (strict): strict own-account implementation iff no mixture replicates the laws at lower cost, or at equal cost with weight on a plan with different laws (250 cases, 102 strict)
PASS tie example: cheapest law-replicating mixture costs exactly the target cost, so only weak implementation is possible
PASS restricting the menu to the target allocations weakly lowers least limited-liability own-account pay, strictly in 37 of 253 environments
PASS economy: 120 feasible allocations; product hull of {0,P,Q} has 29 and contains D and all-B
PASS own-account fee 5/4 per right account: investigation strictly optimal on the whole product hull (3 > 5/2 > 13/10)
PASS own-account fee 5/4: investigation strictly optimal on all 120 allocations too (values {'none': Fraction(5, 2), 'one': Fraction(13, 10), 'both': Fraction(3, 1)})
PASS partial allocation with two holders pays k uninformed (jk/2); empty accounts pay nothing
PASS own-account fee 1 leaves the expert indifferent: incentive cost of separate accounts is 4
PASS own-account fees satisfy the modular identity statewise
PASS the allocation-tied schedule (8 on P/Q when all right, 7 on D) violates modularity: no separate accounts can write it
PASS under that schedule the uninformed recombination D (7) beats investigation (6) on the product hull
PASS joint fee paid only when all accounts are right: least incentive cost also 4
PASS gross CE of the informed policy above outside portfolios: 103.60
PASS uninformed D: gross CE 56.40, net of fee 7 and expense 1: 48.40
PASS net CE: joint mandate with fee 8 and construction 4 = 89.60; own-account fees 5 = 96.60; limit 97.60
PASS statistician: least pay equals c/(1-L) for joint and own-account compensation, and L_joint <= L_own (300 random laws)
PASS diversified edge, four accounts, edge 10 points: own-account pay 6, pay on the joint record 1296/671 = 1.93
PASS edge 5 points: own-account pay 11c; joint pay 1.17c at 20 accounts (ratio 9.37), limit 11
PASS four investors each paying 3 per right account: investigation strictly worthwhile only with all four; each gains 1/5; blind management costs a lone signer 3/2
PASS stand-alone cost: concentrated account 6 > value 2 (no anchor); diversified account 1296/671 < 2 (anchor)
PASS without limited liability a lone concentrated account induces research at expected pay 1 < value 2: anchor
PASS joint pay on four concentrated accounts: each pays 0.483, gains 1.52
PASS Theorem 2: in 200 random formation structures every own-term schedule that implements formation pays the investors it assures more than R* on path; escrow and own constructions at R*+eps implement
PASS equal treatment: a common escrowed payment implements iff it exceeds h-circ; liability m-bar*h-circ >= R* (300 structures)
PASS section 4.7: R* 1.9; common escrowed liability 2.4; common own outlay 3.2 (thousands)
PASS order: with known willingness and public sequential commitments every subgame-perfect equilibrium forms the investment
PASS failure payments leave the formed equilibria unchanged, so assurance raises worst-case welfare by at most the selection gap G
PASS k=3 economy: payments of 3 per right account cost 7.2 in expectation and give each investor 1/5; joint pay 1296/671 gives about 1.52
PASS mixed schedule t1=1.1, beta2=1.1, h=(1,1,1), unanimity: implements formation; own plus escrowed payments to the assured (2.2) exceed R*=2
PASS mixed own and escrowed schedules (288000): implement iff the investors with t+beta>h, plus at most one, form the investment, and then pay or promise more than R* to the assured members of a minimal forming coalition
PASS rounds: in 400 random structures, assignments and schedules, every pure subgame-perfect equilibrium forms the investment iff the assured plus at most one investor per round form it (0 without pure SPE)
PASS rounds: least liability over all assignments and assured sets equals R*_l = min_F(sum h - l largest) (750 cases); zero iff a forming coalition has at most l members; R*_1 = R*
PASS four indispensable investors: least liability 1,900, 1,100, 500 and 0 with one to four rounds
PASS order under private willingness (1712 random economies, payments below expenses): formation exactly when all are willing iff (1-Q_k)(h_k-beta_k) < Q_k x_k for every k < n; expected cost sum P_k(1-Q_k)beta_k
PASS private willingness: payments just above max(0, h_k - Q_k x_k/(1-Q_k)) stay below h_k, implement efficient formation, and cost the stated infimum
PASS two investors, p=0.6, gain 1, expense 2: first mover expects -0.2; efficient formation adds 0.24; escrow interval (0.5, 2); expected cost 0.12
PASS under CARA risk aversion the escrowed payment that induces an early commitment costs less in expectation than the own payment that does
PASS rounds with a fixed assignment: least liability equals min_F(sum h - sum over rounds of the largest expense in F within the round) (200 cases)
PASS private willingness: the order sorted by increasing p_i h_i/(1-p_i) minimizes expected wasted expenses (400 random economies, all orders), and every order wastes strictly less than simultaneous willing formation
PASS two-investor example, simultaneous: payments 2.1 and 0.51 give willing formation as the unique pure Bayesian equilibrium; cost approaches 0.60; waste 0.96 against 0.48 in the order
PASS simultaneous commitments with payments at most the expenses: universal refusal is a Bayesian equilibrium
42 checks; 0 failures
