These Appendixes present the proof of Lemma 0, an extension of our equilibrium characterization allowing for mixed strategies, an analysis of the full commitment case, and details for Example 4 showing that path dependence can arise when shocks are ergodic.KEYWORDS: Principal-agent model, adverse selection, ratchet effect, inefficiency, learning, path dependence. Then there must exist a history (h t b t ) with h t h t that arises on the path of play with positive probability at which the principal offers a transfer T t > c k [ht ] that type c k[ht ] accepts. Note first that since type c k [h t ] accepts offer T t , all types in the support of C[h t ] must also accept it. Indeed, if this were not true, then there would be a highest type c k ∈ C[h t ] that rejects the offer. By the induction hypothesis, the equilibrium payoff that this type obtains at history h t is V (σ μ) k [h t b t ] = 0, since this type would be the highest cost in the support of the principal's beliefs following a rejection. But this cannot be, since type c k can get a payoff of at least T t − c k > 0 by accepting the principal's offer at time t .We now construct an alternative strategy profileσ that is otherwise identical to σ except that at history (h t b t ), the agent is offered a transferT ∈ (c k[ht ] T t ). Specify the principal's beliefs at history (h t b t ) as follows: regardless of the agent's action, the principal's beliefs at the end of the period are the same as her beliefs at the beginning of the period. At all other histories, the principal's actions and beliefs are the same as in the original equilibrium. Note that, given these beliefs, at history h t , all agent types in C[h t ] find it strictly optimal to accept the principal's offerT and take the action. Thus, the principal's payoff at history h t is larger than her payoff under the original equilibrium, which cannot be since the original equilibrium was in Σ K .Q.E.D.
PROOF OF PART (II):The proof is by induction of the cardinality of C[h t ]. Consider first a history h t such that |C[h t ]| = 1. Since |C[h t ]| = 1 < 2, the claim is vacuously true.Avidit Acharya: avidit@stanford.edu Juan Ortner: jortner@bu.edu 1 Indeed, if C[h t ] = {c i }, then in any PBE in Σ K the agent takes action a = 1 at time t ≥ t if and only if b t ∈ E i , and the principal pays the agent a transfer equal to c i every time the agent takes the action.